Information processing device, information processing method, and information processing program

Abstract

To realize an information processing device that can suitably generate a detailed description of symptoms. [Solution] The information processing device includes a determination unit that determines whether the order information group corresponds to a type for which a detailed symptom description needs to be generated; an identification unit that identifies related order information related to the corresponding type if it is determined to correspond; an extraction unit that extracts event-related information related to a medical event from the medical examination record; and a generation unit that generates a detailed symptom description by inputting the related order information and the event-related information into a language model.

Description

Technical Field

[0001] This disclosure relates to an information processing apparatus, an information processing method, and an information processing program.

Background Art

[0002] Techniques for generating medical documents using a language model are known. Patent Document 1 discloses an electronic medical record system that provides the content of an electronic medical record, which extracts information on items necessary for processing documents in a specified category from the electronic medical record, to a large language model and generates an order document or the like.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In the electronic medical record system described in Patent Document 1, a configuration for generating a detailed symptom description is not described. A detailed symptom description is a medical document for explaining medical validity in the case of non-standard medical practices such as overlapping examinations or administering drugs with overlapping effects. There is a need for a technique for generating such a detailed symptom description using a language model.

[0005] This disclosure has been made in view of the above problems, and an exemplary object thereof is to provide a technique for suitably generating a detailed symptom description.

Means for Solving the Problems

[0006] An information processing device relating to an exemplary aspect of this disclosure includes: a determination means for referring to a group of order information containing a plurality of order information indicating each of a plurality of medical instructions, and determining whether the group of order information falls under one or more types for which a detailed description of symptoms needs to be generated; if the determination means determines that the group of order information falls under one or more types for which a detailed description of symptoms needs to be generated, a identification means for identifying related order information from the order information contained in the group of order information that is related to the corresponding type; an extraction means for extracting event-related information from a medical examination record that is related to the related order information and is related to a medical event; and a generation means for generating a detailed description of symptoms by inputting the related order information and the event-related information into a language model.

[0007] An information processing method relating to an exemplary aspect of this disclosure includes: a determination process in which at least one processor refers to a group of order information including a plurality of order information indicating each of a plurality of medical instructions and determines whether the group of order information falls under one or more types for which a detailed symptom description needs to be generated; if, in the determination process, the group of order information is determined to fall under one or more types for which a detailed symptom description needs to be generated, the at least one processor identifies related order information from the order information included in the group of order information that is related to the corresponding type; an extraction process in which the at least one processor extracts event-related information from a medical examination record that is related to the related order information and is related to a medical event; and a generation process in which the at least one processor generates a detailed symptom description by inputting the related order information and the event-related information into a language model.

[0008] An illustrative aspect of the present disclosure is an information processing program that causes a computer to function as an information processing device, the program causing the computer to function as: a determination means that refers to a group of order information including a plurality of order information indicating each of a plurality of medical instructions and determines whether the group of order information falls under one or more types for which the generation of a detailed symptom description is required; if the determination means determines that the group of order information falls under one or more types for which the generation of a detailed symptom description is required, a identification means that identifies related order information from the order information included in the group of order information that is related to the corresponding type; an extraction means that extracts event-related information from a medical record that is related to the related order information and is related to a medical event; and a generation means that generates a detailed symptom description by inputting the related order information and the event-related information into a language model. [Effects of the Invention]

[0009] According to an illustrative aspect of this disclosure, one exemplary effect is that it can provide a technique for suitably generating detailed symptom descriptions. [Brief explanation of the drawing]

[0010] [Figure 1] This is a block diagram showing the configuration of the information processing device related to this disclosure. [Figure 2] This is a flowchart showing the flow of the information processing method related to this disclosure. [Figure 3] This is a block diagram showing the configuration of the information processing device related to this disclosure. [Figure 4] This is a flowchart showing the flow of the information processing method related to this disclosure. [Figure 5] This figure shows an example of an image output by the output unit related to this disclosure. [Figure 6] This figure shows another example of an image output by the output unit relating to this disclosure. [Figure 7] This figure shows yet another example of an image output by the output unit relating to this disclosure. [Figure 8]This is a block diagram showing the configuration of a computer that functions as an information processing device related to this disclosure. [Modes for carrying out the invention]

[0011] The following are examples of embodiments of the present invention. However, the present invention is not limited to the exemplary embodiments shown below, and various modifications are possible within the scope of the claims. For example, embodiments obtained by appropriately combining some or all of the technologies (things or methods) employed in each of the exemplary embodiments shown below may also be included in the scope of the present invention. Furthermore, embodiments obtained by appropriately omitting some of the technologies employed in each of the exemplary embodiments shown below may also be included in the scope of the present invention. In addition, the effects mentioned in each of the exemplary embodiments shown below are examples of effects that can be expected in that exemplary embodiment and do not define the scope of the present invention. That is, embodiments that do not produce the effects mentioned in each of the exemplary embodiments shown below may also be included in the scope of the present invention.

[0012] [First Exemplary Embodiment] A first exemplary embodiment, which is an example of an embodiment of the present invention, will be described in detail with reference to the drawings. This exemplary embodiment is the basic form for each of the exemplary embodiments described later. The scope of application of each technology adopted in this exemplary embodiment is not limited to this exemplary embodiment. That is, each technology adopted in this exemplary embodiment can also be adopted in other exemplary embodiments included in this disclosure, to the extent that no particular technical problems occur. Furthermore, each technology shown in the drawings referenced to explain this exemplary embodiment can also be adopted in other exemplary embodiments included in this disclosure, to the extent that no particular technical problems occur.

[0013] (Configuration of Information Processing Device 1) The configuration of the information processing apparatus 1 will be described with reference to FIG. 1. FIG. 1 is a block diagram showing the configuration of the information processing apparatus 1. As shown in FIG. 1, the information processing apparatus 1 includes a determination unit 11, a specification unit 12, an extraction unit 13, and a generation unit 14. The determination unit 11, the specification unit 12, the extraction unit 13, and the generation unit 14 respectively realize determination means, specification means, extraction means, and generation means in this exemplary embodiment.

[0014] (Determination Unit 11) The determination unit 11 refers to an order information group including a plurality of order information indicating each of a plurality of medical instructions, and determines whether the order information group corresponds to any one of one or more types for which generation of symptom details is necessary. The determination unit 11 supplies the determination result to the specification unit 12.

[0015] The determination unit 11 may be configured to refer to a receipt information group including a plurality of receipt information indicating each of a plurality of medical fee statements in addition to the order information group.

[0016] (Specification Unit 12) When the determination unit 11 determines that the order information group corresponds to any one of one or more types for which generation of symptom details is necessary, the specification unit 12 specifies related order information related to the corresponding type among the order information included in the order information group. The specification unit 12 supplies the specified related order information to the extraction unit 13 and the generation unit 14.

[0017] (Extraction Unit 13) The extraction unit 13 extracts event-related information related to a medical event, which is information related to the related order information, from the examination record. The extraction unit 13 supplies the extracted event-related information to the generation unit 14.

[0018] (Generation Unit 14) The generation unit 14 generates symptom details by inputting the related order information and the event-related information into a language model.

[0019] (Effect of Information Processing Apparatus 1) As described above, the information processing device 1 employs a configuration comprising: a determination unit 11 that refers to a group of order information containing multiple order information indicating each of multiple medical instructions and determines whether the group of order information falls under one or more types for which a detailed symptom description needs to be generated; a determination unit 12 that, if the determination unit 11 determines that the group of order information falls under one or more types for which a detailed symptom description needs to be generated, identifies related order information from the order information contained in the group that is related to the corresponding type; an extraction unit 13 that extracts event-related information from the medical examination record that is related to the related order information and is related to a medical event; and a generation unit 14 that generates a detailed symptom description by inputting the related order information and the event-related information into a language model.

[0020] Therefore, the information processing device 1 identifies the relevant order information necessary for generating the symptom description from the order information group for which the symptom description needs to be generated. Furthermore, the information processing device 1 extracts event-related information related to the relevant order information from the medical examination report, and generates the symptom description by inputting the relevant order information and event-related information into the language model. In other words, when a symptom description is required, the information processing device 1 identifies the information necessary for generating the symptom description from the order information group and the medical examination report, and generates the symptom description using the language model. Thus, the information processing device 1 can suitably generate the symptom description.

[0021] (Information processing method S1 flow) The flow of the information processing method S1 will be explained with reference to Figure 2. Figure 2 is a flowchart showing the flow of the information processing method S1. As shown in Figure 2, the information processing method S1 includes a determination process S11, a specific process S12, an extraction process S13, and a generation process S14.

[0022] (Decision process S11) In the determination process S11, the determination unit 11 refers to a group of order information containing multiple order information indicating each of multiple medical instructions, and determines whether the group of order information corresponds to one or more types for which a detailed symptom description needs to be generated. The determination unit 11 supplies the determination result to the identification unit 12.

[0023] (Specific processing S12) In the identification process S12, if the identification unit 12 determines in the determination process S11 that the order information group corresponds to one or more types for which a detailed symptom description needs to be generated, it identifies related order information from the order information group that is related to the corresponding type. The identification unit 12 supplies the identified related order information to the extraction unit 13 and the generation unit 14.

[0024] (Extraction part 13) In extraction process S13, the extraction unit 13 extracts event-related information from the medical examination report that is related to the relevant order information and is related to the medical event. The extraction unit 13 supplies the extracted event-related information to the generation unit 14.

[0025] (Generation unit 14) In generation process S14, the generation unit 14 generates a detailed symptom description by inputting related order information and event-related information into the language model.

[0026] (Effects of information processing method S1) As described above, the information processing method S1 employs a configuration that includes: a determination unit 11 which refers to a group of order information containing multiple order information indicating each of multiple medical instructions and determines whether the group of order information falls under one or more types for which a detailed symptom description needs to be generated; a determination process S11 in which the identification unit 12, if determined in the determination process S11, determines that the group of order information falls under one or more types for which a detailed symptom description needs to be generated, identifies related order information from the order information contained in the group that is related to the corresponding type; an extraction unit 13 which extracts event-related information from the medical examination record that is related to the related order information and is related to a medical event; and a generation unit 14 which generates a detailed symptom description by inputting the related order information and event-related information into a language model. Therefore, the same effects as those of the information processing device 1 described above can be obtained with the information processing method S1.

[0027] [Second exemplary embodiment] A second exemplary embodiment, which is an example of an embodiment of the present invention, will be described in detail with reference to the drawings. Components having the same function as those described in the above-described exemplary embodiment are denoted by the same reference numerals, and their descriptions are omitted as appropriate. The scope of application of each technology adopted in this exemplary embodiment is not limited to this exemplary embodiment. That is, each technology adopted in this exemplary embodiment can also be adopted in other exemplary embodiments included in this disclosure, to the extent that no particular technical problems arise. Furthermore, each technology shown in the drawings referenced to describe this exemplary embodiment can also be adopted in other exemplary embodiments included in this disclosure, to the extent that no particular technical problems arise.

[0028] (Overview of Information Processing Device 1A) The information processing device 1A determines whether or not it is necessary to generate a symptom detail SD, and if it is determined that it is necessary, it generates a symptom detail SD using the language model LM.

[0029] More specifically, the information processing device 1A first refers to a group of order information OIG containing multiple order information OIs, each representing a different medical instruction, and determines whether the group of order information OIGs falls under one or more types for which the generation of a symptom detail SD is required.

[0030] Examples of medical instructions indicated by order information (OI) include, but are not limited to, instructions for tests, medication, surgery, and blood transfusions. Furthermore, the number of order information (OI) included in the order information group (OIG) is not particularly limited; it may include multiple order information (OI) relating to medical procedures for a single patient, or multiple order information (OI) relating to medical procedures for a specific group of patients.

[0031] The types of medical procedures that require the generation of a detailed symptom-based medical record (SD) are those that are non-standard medical procedures requiring a detailed symptom-based SD. In this disclosure, examples of procedures that require the generation of a detailed symptom-based SD include, but are not limited to, "duplicate testing" (where tests are performed multiple times on the same patient), "duplicate medication" (where medications with overlapping effects are prescribed to the same patient), "high-cost surgery" (where the surgical procedure is expensive), and "blood transfusion."

[0032] In other words, an order information group OIG is considered to fall under one or more types requiring the generation of a symptom-detailing SD if each of the multiple order information OIs included in the order information group OIG indicates a "duplicate test," "duplicate medication," "expensive surgery," or "blood transfusion."

[0033] If the order information group OIG is determined to be one or more types for which the generation of a symptom detail SD is required, the information processing device 1A identifies the related order information ROI from the order information OI contained in the order information group OIG that is associated with the corresponding type.

[0034] For example, if, among the multiple order information OIs included in the order information group OIG, two drugs with overlapping efficacy (drug A and drug B) were administered to a single patient, the information processing device 1A determines that the order information group OIG corresponds to "duplicate medication". Next, the information processing device 1A identifies related order information ROIs (related order information ROI_A indicating that drug A was administered, and related order information ROI_B indicating that drug B was administered) from the multiple order information OIs that are related to "duplicate medication".

[0035] Next, the information processing device 1A extracts event-related information (ERI), which is related to the medical event, from the medical examination article information (MEI), which represents the medical examination article.

[0036] A medical examination report is information that describes what a doctor did when examining a patient. For example, a medical examination report is information recorded in an electronic medical record. More specifically, a medical examination report includes the patient's chief complaint as stated by the doctor, the patient's medical history (past and present conditions), the doctor's examination findings, the treatment plan, and the treatment outcome.

[0037] Event-related information (ERI) refers to information related to medical events affecting a patient, such as consultations, tests, surgeries, and medications. Examples of event-related information (ERI) include detailed patient symptoms, detailed medical procedures, and detailed outcomes.

[0038] The information processing device 1A then generates a symptom description SD by inputting the identified related order information ROI and the extracted event-related information ERI into the language model LM.

[0039] Specific examples of the processes performed by the information processing device 1A will be described later.

[0040] (Configuration of Information Processing Device 1A) The configuration of the information processing device 1A will be described with reference to Figure 3. Figure 3 is a block diagram showing the configuration of the information processing device 1A. As shown in Figure 3, the information processing device 1A includes a control unit 10, a storage unit 20, an input / output unit 21, and a communication unit 22.

[0041] (Storage unit 20) The memory unit 20 stores data that the control unit 10 references. As an example, as shown in Figure 3, the memory unit 20 stores an order information group OIG containing multiple order information OIs, medical record information MEI, a first machine learning model TM1, a second machine learning model TM2, a third machine learning model TM3, and a language model LM. When we say that the first machine learning model TM1, the second machine learning model TM2, the third machine learning model TM3, and the language model LM are stored in the memory unit 20, it means that the parameters defining each of the first machine learning model TM1, the second machine learning model TM2, the third machine learning model TM3, and the language model LM are stored in the memory unit 20.

[0042] The order information OI, order information group OIG, and medical examination article information MEI are as described above.

[0043] (First machine learning model TM1) The first machine learning model TM1 is a machine learning model that takes order information group OIG as input and outputs a determination result that determines whether or not the order information group OIG corresponds to one or more types for which the generation of symptom detail SD is required. As an example, the first machine learning model TM1 is a machine learning model that has been trained using training data that includes multiple pairs of one or more order information OI and information indicating whether or not the one or more order information OI corresponds to a type.

[0044] Alternatively, the first machine learning model TM1 may be a machine learning model that has been trained to take an order information set OIG as input and output type information TY that indicates the type to which the order information set OIG corresponds. As an example in this case, the first machine learning model TM1 is a machine learning model that has been trained using training data that includes multiple pairs of one or more order information OI and information indicating the type to which the one or more order information OI corresponds.

[0045] Alternatively, the first machine learning model TM1 may be a machine learning model that has been trained to take order information group OIG and type information TY as input and output related order information ROI that is associated with the type indicated by type information TY from among the order information included in order information group OIG. As an example in this case, the first machine learning model TM1 is a machine learning model that has been trained using training data that includes multiple pairs of order information OI and type information TY corresponding to the one or more order information OI, and related order information ROI associated with the type indicated by type information TY.

[0046] Here, the first machine learning model TM1 may be composed of multiple machine learning models (machine learning model A to machine learning model C).

[0047] In this case, machine learning model A is a machine learning model that has been trained to take order information set OIG as input and output a determination result that determines whether or not order information set OIG corresponds to one or more types for which the generation of symptom detail SD is required.

[0048] Furthermore, machine learning model B is a machine learning model that takes order information set OIG as input and outputs type information TY, which indicates the type to which order information set OIG corresponds.

[0049] Furthermore, machine learning model C is a machine learning model that takes order information set OIG and type information TY as input and outputs related order information ROI that is associated with the type indicated by type information TY from the order information set OIG.

[0050] (Second machine learning model TM2) The second machine learning model TM2 is a machine learning model that takes a medical consultation article as input and outputs words related to the medical events contained in that consultation article. As an example, the second machine learning model TM2 is a machine learning model that was trained using training data that includes multiple pairs of medical consultation article information MEI and words related to the medical events contained in the consultation article indicated by the medical consultation article information MEI.

[0051] (Third machine learning model TM3) The third machine learning model, TM3, is a machine learning model that takes words related to medical events and related order information ROI as input and outputs an embedding vector of the input words and an embedding vector of medical language resources related to the related order information ROI.

[0052] As an example, the third machine learning model TM3 is a machine learning model trained using multiple training datasets, each consisting of a set of text related to a medical event, at least one of the terms symptoms, treatments, and outcomes; related order information ROI; medical language resources related to the related order information ROI; embedding vectors for at least one of the terms symptoms, treatments, and outcomes; and embedding vectors for the medical language resources.

[0053] As an example of this configuration, the third machine learning model TM3 first identifies medical language resources related to the input relevant order information ROI. Medical language resources are resources of language (terminology) used in medicine. Examples of medical language resources include, but are not limited to, symptoms, disease names, drug names, and drug efficacy.

[0054] The medical language resources associated with the input ROI of relevant order information are those related to the medical instructions indicated by the input ROI of relevant order information. For example, if the input ROI of relevant order information is an instruction to administer drug A, the second machine learning model TM2 will identify drug A, the symptoms that drug A improves, and the names of diseases for which drug A is effective as medical language resources associated with the input ROI of relevant order information.

[0055] Next, the second machine learning model TM2 converts the input text into embedding vectors. The second machine learning model TM2 also converts the identified medical language resources into embedding vectors.

[0056] (Language Model LM) The language model (LM) is a machine learning model trained to take prompts as input and output symptom detail SDs. For example, the language model (LM) is a machine learning model trained using training data containing multiple pairs of prompts and corresponding symptom detail SDs. Examples of prompts include instructions indicated by relevant order information and symptoms, treatments, and outcomes indicated by relevant event information, which are converted into a predetermined format.

[0057] (Input / output section 21) The input / output unit 21 is an interface to input devices that accept data input and output devices that output data. Examples of input devices include, but are not limited to, microphones, cameras, eye-tracking devices, keyboards, and touchpads. Examples of output devices include, but are not limited to, speakers and liquid crystal displays.

[0058] (Communications Section 22) The communication unit 22 is an interface for sending and receiving data over a network. Examples of the communication unit 22 include, but are not limited to, communication chips in various communication standards such as Ethernet®, Wi-Fi®, and wireless communication standards for mobile data communication networks, as well as USB-compliant connectors.

[0059] (Control Unit 10) The control unit 10 controls each component of the information processing device 1A. As shown in Figure 3, the control unit 10 also includes a determination unit 11, a specification unit 12, an extraction unit 13, a generation unit 14, an output unit 15, and an acquisition unit 16. In this exemplary embodiment, the determination unit 11, specification unit 12, extraction unit 13, generation unit 14, and output unit 15 implement the determination means, specification means, extraction means, generation means, and output means, respectively.

[0060] (Judgment section 11) The determination unit 11 refers to the order information group OIG and determines whether the order information group OIG corresponds to one or more types for which the generation of a symptom detail SD is required. The determination unit 11 supplies the determination result to the identification unit 12.

[0061] As an example, the determination unit 11 inputs the order information OIG stored in the memory unit 20 to the first machine learning model TM1, refers to the determination result output from the first machine learning model TM1, and determines whether the order information group OIG corresponds to one or more types for which the generation of a symptom detail SD is required.

[0062] (Specific part 12) If the determination result supplied by the determination unit 11 indicates that the order information group OIG corresponds to one or more types for which the generation of a symptom detail SD is required, the identification unit 12 identifies the related order information ROI from the order information OI included in the order information group OIG that is associated with the corresponding type. The identification unit 12 supplies the identified related order information ROI to the extraction unit 13.

[0063] As an example, the identification unit 12 inputs the order information group OIG into the first machine learning model TM1 and obtains type information TY that indicates the type to which the order information group OIG corresponds. Next, the identification unit 12 inputs the order information group OIG and the obtained type information TY into the first machine learning model TM1 and obtains related order information ROI associated with the type indicated by the type information TY.

[0064] (Extraction part 13) The extraction unit 13 extracts event-related information (ERI) from the medical examination report, which is related to the relevant order information (ROI) and the medical event. The extraction unit 13 supplies the extracted event-related information (ERI) to the generation unit 14.

[0065] As an example, the extraction unit 13 refers to the similarity between the wording contained in the medical examination article that is related to the medical event and the medical language resources related to the related order information ROI, and extracts the event-related information ERI.

[0066] As an example of this configuration, the extraction unit 13 first uses the second machine learning model TM2 to obtain words related to medical events included in the medical consultation article. More specifically, the extraction unit 13 inputs the medical consultation article information MEI into the second machine learning model TM2 to obtain words related to symptoms, treatments, and outcomes.

[0067] The Medical Consultation Report Information (MEI) contains unstructured text information, including unstructured data that includes at least one of the terms for symptoms, treatment, and outcome. The second machine learning model TM2 outputs at least one of the terms for symptoms, treatment, and outcome by generating entities for each medical event based on the unstructured data contained in the Medical Consultation Report Information (MEI).

[0068] Alternatively, the second machine learning model TM2 may be configured to output a timeline in which entities are arranged chronologically. In this case, the second machine learning model TM2 extracts medical named entities and their temporal relationships from unstructured text data and outputs a group of unstructured data entities whose temporal relationships are defined based on the order in which medical events occur.

[0069] Next, the extraction unit 13 uses a third machine learning model to reference the similarity between the embedding vectors of words related to medical events and the embedding vectors of medical language resources, and extracts event-related information (ERI).

[0070] Specifically, the extraction unit 13 inputs the text obtained from the second machine learning model TM2 and the related order information ROI into the third machine learning model TM3, and obtains embedding vectors of at least one of the symptoms, treatments, and outcomes indicated by the input information, and embedding vectors of medical language resources related to the related order information ROI.

[0071] The extraction unit 13 then calculates the similarity between the embedding vector of at least one of the words for symptoms, treatments, and outcomes and the embedding vector of the medical language resource related to the relevant order information ROI. For example, the extraction unit 13 calculates the cosine similarity between the embedding vector of at least one of the words for symptoms, treatments, and outcomes and the embedding vector of the medical language resource related to the relevant order information ROI as the similarity.

[0072] In this configuration, the extraction unit 13 may sort the symptom, treatment, and outcome phrases in descending order of similarity (or descending order of similarity), extract event-related information (ERI) representing the symptom, treatment, and outcome phrases, and supply it to the generation unit 14. Here, the extraction unit 13 may be configured to sort only the symptom, treatment, and outcome phrases whose similarity is higher than a predetermined value.

[0073] In this way, the extraction unit 13 refers to the similarity between the wording related to the medical event and the medical language resources related to the related order information ROI, and extracts the event-related information. Therefore, the extraction unit 13 can suitably extract even wording that is inconsistently written in the medical examination report.

[0074] Furthermore, the extraction unit 13 sorts the symptom, treatment, and outcome phrases in descending order of similarity (or descending order of similarity), and then extracts event-related information (ERIs) representing the symptom, treatment, and outcome phrases, thereby notifying which phrases have a high probability of being event-related information (ERIs).

[0075] (Generation unit 14) The generation unit 14 generates a symptom description SD by inputting the related order information ROI and event-related information supplied from the extraction unit 13 into the language model LM. The generation unit 14 supplies the generated symptom description SD to the output unit 15.

[0076] As an example, the generation unit 14 first generates prompts by converting the instructions indicated by the relevant order information and the symptom, treatment, and outcome texts indicated by the event-related information supplied by the extraction unit 13 into a predetermined format. Then, the generation unit 14 inputs the generated prompts into the language model LM. The generation unit 14 then obtains the symptom detail SD output from the language model LM.

[0077] For example, consider the following scenario for the relevant order information ROI and the wording for symptoms, treatments, and outcomes: Related order information: "Administer drug A", "Administer drug B" Symptoms: "Eosinophilic granulomatous vasculitis" Treatment: "Administer drug A," "Administer drug B" • Outcome: "Frequency of seizures decreased after starting medication B." In this case, the generation unit 14 generates a prompt that reads, "When drug A and drug B were administered to a patient exhibiting symptoms of eosinophilic granulomatous vasculitis, the frequency of attacks decreased after starting drug B. Please generate a detailed description of the symptoms after administering drug A and drug B."

[0078] Furthermore, if there are multiple event-related pieces of information supplied from the extraction unit 13, the generation unit 14 may be configured to allow the user to select the event-related information to be input into the language model LM. An example of this configuration will be described later.

[0079] (Output section 15) The output unit 15 outputs data to the input / output unit 21 or the communication unit 22. As an example, the output unit 15 outputs the symptom details SD generated by the generation unit 14. As another example, the output unit 15 outputs event-related information ERI. Examples of images output by the output unit 15 will be described later.

[0080] (Acquisition part 16) The acquisition unit 16 acquires input information from the input / output unit 21 that indicates user input. For example, the acquisition unit 16 acquires input information that indicates the item selected by the user. Examples of information acquired by the acquisition unit 16 will be described later.

[0081] (Example of processing performed by the information processing device 1A) An example of a process (information processing method S1A) executed by the information processing device 1A will be explained with reference to Figure 4. Figure 4 is a flowchart showing the flow of information processing method S1A.

[0082] (Decision process S11) In the determination process S11, the determination unit 11 refers to the order information group OIG and determines whether the order information group OIG corresponds to one or more types for which the generation of a symptom detail SD is required. Specifically, it executes the following processing steps S111 and S112.

[0083] (Step S111) In step S111, the determination unit 11 inputs the order information OIG stored in the memory unit 20 to the first machine learning model TM1.

[0084] (Step S112) In step S112, the determination unit 11 refers to the determination result output from the first machine learning model TM1 and determines whether the order information group OIG corresponds to one or more types for which the generation of a symptom detail SD is required. The determination unit 11 supplies the determination result to the identification unit 12.

[0085] In step S112, if it is determined that the order information group OIG does not fall under any of the one or more types for which the generation of a symptom detail SD is required (step S112: NO), the information processing device 1A terminates the information processing method S1A shown in Figure 4.

[0086] (Specific processing S12) In step S112, if it is determined that the order information group OIG corresponds to one or more types for which the generation of a symptom detail SD is required (step S112: YES), in the identification process S12, the identification unit 12 identifies related order information ROIs from the order information OIs included in the order information group OIG that are related to the corresponding type. The identification unit 12 supplies the identified related order information ROIs to the extraction unit 13.

[0087] As an example, the identification unit 12 inputs the order information group OIG into the first machine learning model TM1 and obtains type information TY that indicates the type to which the order information group OIG corresponds. Next, the identification unit 12 inputs the order information group OIG and the obtained type information TY into the first machine learning model TM1 and obtains related order information ROI associated with the type indicated by the type information TY.

[0088] (Extraction process S13) In extraction process S13, the extraction unit 13 extracts event-related information (ERI) from the medical examination report, which is related to the relevant order information (ROI) and related to the medical event. Specifically, the extraction unit 13 executes the following steps S131 to S134.

[0089] (Step S131) In step S131, the extraction unit 13 inputs the medical examination article information MEI into the second machine learning model TM2 to obtain at least one of the following words: symptom, treatment, and outcome.

[0090] (Step S132) In step S132, the extraction unit 13 inputs the text obtained from the second machine learning model TM2 and the related order information ROI into the third machine learning model TM3, and obtains embedding vectors of at least one of the symptoms, treatments, and outcomes indicated by the input information, and embedding vectors of medical language resources related to the related order information ROI.

[0091] (Step S133) In step S133, the extraction unit 13 calculates the similarity between the embedding vector of at least one of the words for symptoms, treatments, and outcomes and the embedding vector of the medical language resource related to the relevant order information ROI.

[0092] (Step S134) In step S134, the extraction unit 13 sorts the symptom, treatment, and outcome phrases in descending order of similarity (or descending order of similarity), and then supplies event-related information (ERI) representing the symptom, treatment, and outcome phrases to the generation unit 14.

[0093] (Generation process S14) In the generation process S14, the generation unit 14 generates a symptom description SD by inputting the related order information ROI and the event-related information ERI supplied from the extraction unit 13 into the language model LM. Specifically, the generation unit 14 executes the following steps S141 to S143.

[0094] (Step S141) In step S141, the generation unit 14 generates prompts that convert the instructions indicated by the relevant order information and the symptom, treatment, and outcome texts indicated by the event-related information supplied from the extraction unit 13 into a predetermined format.

[0095] (Step S142) In step S142, the generation unit 14 generates a symptom description SD by inputting the generated prompt into the language model LM.

[0096] (Output processing S15) In output processing S15, the output unit 15 outputs the generated symptom detail SD.

[0097] (Example 1 of an image output by output unit 15) An example of an image output by the output unit 15 will be explained with reference to Figure 5. Figure 5 is a diagram showing an example of an image output by the output unit 15.

[0098] If there are multiple order information groups OIG, the output unit 15 may output an image to ask the user which order information group OIG to use before the determination process S11 is executed.

[0099] For example, if there are separate order information groups (OIGs) for multiple patients, the output unit 15 outputs an image asking the user which patient's order information group (OIG) to use, as shown in Figure 5. Alternatively, as shown in Figure 5, the output unit 15 may also output an image including a status indicating whether or not a symptom detail SD has been generated.

[0100] If the acquisition unit 16 acquires input information indicating any patient from the image shown in Figure 5, the determination unit 11 acquires the patient order information group OIG indicated by the input information acquired by the acquisition unit 16 in the determination process S11. The determination unit 11 then determines whether the acquired order information group OIG corresponds to one or more types for which the generation of a symptom detail SD is required.

[0101] For example, if the acquisition unit 16 acquires input information indicating that the user has selected patient ID "0000001" for the image shown in Figure 5, the determination unit 11 acquires the order information group OIG for patient ID "0000001" in the determination process S11. The determination unit 11 then determines whether the acquired order information group OIG for patient ID "0000001" corresponds to one or more types for which the generation of a symptom detail SD is required.

[0102] With this configuration, the information processing device 1A can identify the order information group OIG for the user to generate a symptom detail SD.

[0103] (Example 2 of an image output by output unit 15) Other examples of images output by the output unit 15 will be explained with reference to Figure 6. Figure 6 is a diagram showing other examples of images output by the output unit 15.

[0104] The output unit 15 may output an image including medical examination record information MEI, type information TY, and event-related information ERI.

[0105] For example, the output unit 15 outputs an image containing medical record information (MEI) as shown on the left side of Figure 6.

[0106] Furthermore, the output unit 15 outputs an image containing type information TY1 to TY3, and event-related information ERI1 to ERI5, as shown on the right side of Figure 6. The output unit 15 may also display the event-related information ERI1 to ERI5, etc., in descending order of similarity (or in descending order of similarity). With this configuration, the extracted event-related information ERI1 to ERI5, etc., can be notified to the user.

[0107] Furthermore, the output unit 15 may output an image containing the language model LM to be used, as shown in the lower right of Figure 6.

[0108] Furthermore, if the input unit 16 acquires user input information for the image shown in Figure 6, the information processing device 1A may perform processing based on the acquired input information.

[0109] For example, if the acquisition unit 16 acquires input information indicating that the user has selected type information TY1 "duplicate medication" for the image shown in Figure 6, the generation unit 14 generates a symptom detail SD in generation process S14 by inputting related order information ROI and event-related information ERI1 to ERI5 into the language model LM.

[0110] As another example, if the acquisition unit 16 acquires input information indicating that the user will delete "headache medicine 1" from event-related information ERI1 for the image shown in Figure 6, the generation unit 14 generates a symptom detail SD in generation process S14 by inputting the related order information ROI and event-related information ERI2 to ERI5 into the language model LM.

[0111] As another example, if the acquisition unit 16 acquires input information indicating that the user wants to change "headache medicine 2" to "headache medicine 3" in the event-related information ERI3 for the image shown in Figure 6, the generation unit 14, in generation process S14, changes the event-related information ERI3 from "headache medicine 2" to "headache medicine 3," and then inputs the related order information ROI and event-related information ERI2 to ERI5 into the language model LM to generate a symptom detail SD.

[0112] As another example, if the acquisition unit 16 acquires input information indicating that the user selected "hearing aid adjustment" included in the medical examination article information MEI for the image shown in Figure 6, the extraction unit 13 extracts "hearing aid adjustment" as event-related information ERI in the extraction process S13.

[0113] As another example, if the acquisition unit 16 acquires input information indicating that the language model LM to be used should be changed from "GPT3.5" to "GPT4" as shown in Figure 6, the generation unit 14 will use "GPT4" as the language model LM in the generation process S14.

[0114] (Example 3 of an image output by output unit 15) Further examples of images output by the output unit 15 will be explained with reference to Figure 7. Figure 7 is a diagram showing yet another example of an image output by the output unit 15.

[0115] The output unit 15 may output an image that includes the generated symptom description SD. For example, the output unit 15 outputs an image in which the symptom description SD is superimposed on the image in Figure 6.

[0116] Furthermore, if the input unit 16 acquires user input information for the image shown in Figure 7, the information processing device 1A may perform processing based on the acquired input information.

[0117] For example, if the acquisition unit 16 acquires input information indicating that the user wants to generate a symptom detail SD again for the image shown in Figure 7, the generation unit 14 generates the symptom detail SD again in generation process S14.

[0118] As another example, if the input unit 16 acquires input information indicating that the user has modified the symptom detail SD for the image shown in Figure 7, the output unit 15 will display the modified symptom detail SD.

[0119] Furthermore, as shown in the image in Figure 7, a symptom detail SD modified by the user may be stored in a database (e.g., storage unit 20) linked to the patient ID, for example, when the information processing device 1A obtains a document that requires the inclusion of a symptom detail SD (e.g., a medical claim form, a patient's medical examination report, etc.), it obtains the symptom detail SD and other necessary items from the database, automatically creates the document, and outputs it. As another example, the information processing device 1A obtains the format of a document that requires the inclusion of a symptom detail SD, automatically transcribes the symptom detail SD and other necessary items that the user has confirmed, modified, and approved, automatically creates the document, and outputs it.

[0120] (Effects of Information Processing Device 1A) As described above, the information processing device 1A uses the first machine learning model TM1 (or machine learning model A) to determine whether the order information group OIG corresponds to one or more types for which the generation of a symptom detail SD is necessary. Therefore, the information processing device 1A can suitably determine whether the order information group OIG corresponds to one or more types for which the generation of a symptom detail SD is necessary.

[0121] Furthermore, the information processing device 1A uses the first machine learning model TM1 (or machine learning model B) to identify type information TY that indicates the type to which the order information group OIG corresponds. Therefore, the information processing device 1A can suitably identify type information TY that indicates the type to which the order information group OIG corresponds.

[0122] Furthermore, the information processing device 1A identifies related order information ROI using the first machine learning model TM1 (or machine learning model C). Therefore, the information processing device 1A can suitably identify related order information ROI.

[0123] Furthermore, the information processing device 1A uses a second machine learning model TM2 to extract words related to medical events from the medical examination report. The information processing device 1A also uses a third machine learning model TM3 to extract embedding vectors of words related to medical events and embedding vectors of medical language resources related to the associated order information ROI. The information processing device 1A then refers to the similarity between the embedding vectors of words related to medical events and the embedding vectors of medical language resources to extract event-related information ERI. Therefore, the information processing device 1A can suitably extract event-related information ERI even if there are variations in wording in the medical examination report.

[0124] [Examples of implementation using software] Some or all of the functions of the information processing devices 1 and 1A (hereinafter also referred to as "the above devices") may be implemented by hardware such as integrated circuits (IC chips) or by software.

[0125] In the latter case, each of the above devices is implemented, for example, by a computer that executes instructions for a program, which is software that realizes each function. An example of such a computer (hereinafter referred to as Computer C) is shown in Figure 8. Figure 8 is a block diagram showing the hardware configuration of Computer C, which functions as each of the above devices.

[0126] Computer C comprises at least one processor C1 and at least one memory C2. Memory C2 stores a program P that causes computer C to operate as each of the above-mentioned devices. In computer C, processor C1 reads program P from memory C2 and executes it, thereby realizing each of the above-mentioned devices.

[0127] For processor C1, for example, a CPU (Central Processing Unit), GPU (Graphic Processing Unit), DSP (Digital Signal Processor), MPU (Micro Processing Unit), FPU (Floating Point Number Processing Unit), PPU (Physics Processing Unit), TPU (Tensor Processing Unit), quantum processor, microcontroller, or a combination thereof can be used. For memory C2, for example, flash memory, HDD (Hard Disk Drive), SSD (Solid State Drive), or a combination thereof can be used.

[0128] Computer C may also be equipped with RAM (Random Access Memory) for loading program P at runtime and for temporarily storing various data. Furthermore, computer C may be equipped with communication interfaces for sending and receiving data with other devices. Additionally, computer C may be equipped with input / output interfaces for connecting input / output devices such as keyboards, mice, displays, and printers.

[0129] Furthermore, program P can be recorded on a non-temporary, tangible recording medium M that is readable by computer C. Such a recording medium M could be, for example, tape, disk, card, semiconductor memory, or programmable logic circuitry. Computer C can acquire program P via such a recording medium M. Program P can also be transmitted via a transmission medium. Such a transmission medium could be, for example, a communication network or broadcast waves. Computer C can also acquire program P via such a transmission medium.

[0130] Furthermore, each of the above functions of each of the above devices may be implemented by a single processor in a single computer, by multiple processors in a single computer working together, or by multiple processors in each of multiple computers working together. In addition, the programs for implementing each of the above functions in each of the above devices may be stored in a single memory in a single computer, distributed and stored in multiple memories in a single computer, or distributed and stored in multiple memories in each of multiple computers.

[0131] [Additional Note A] This disclosure includes the technologies described in the following appendices. However, the present invention is not limited to the technologies described in the following appendices, and various modifications are possible within the scope of the claims.

[0132] (Note A1) A determination means that refers to a group of order information containing multiple order information indicating each of multiple medical instructions, and determines whether the group of order information corresponds to one or more types for which a detailed description of symptoms needs to be generated, If the determination means determines that the order information group corresponds to one or more types for which a detailed description of symptoms is required, the determination means identifies related order information from the order information group that is associated with the corresponding type, An extraction means for extracting event-related information related to the medical event, which is related to the aforementioned related order information, from the medical examination article, A generation means for generating a detailed description of symptoms by inputting the aforementioned related order information and the aforementioned event-related information into a language model, An information processing device equipped with the following features.

[0133] (Appendix A2) The determination means takes the order information set as input and uses a first machine learning model trained to output a determination result that determines whether the order information set corresponds to one or more types for which a detailed symptom description needs to be generated, to determine whether the order information set corresponds to one or more types for which a detailed symptom description needs to be generated. The information processing device described in Appendix A1.

[0134] (Note A3) The first machine learning model is further trained to take the order information set as input and output type information indicating the type to which the order information set belongs, if it determines that the order information set belongs to one or more types to which a detailed description of symptoms is required. The information processing device described in Appendix A2.

[0135] (Note A4) The first machine learning model is further trained to take the order information set and the type information as input and output related order information from the order information set that is related to the type indicated by the type information, The identification means identifies the related order information by inputting the order information set and the type information into the first machine learning model. The information processing device described in Appendix A3.

[0136] (Note A5) The extraction means extracts event-related information by referring to the similarity between the wording contained in the medical examination article that relates to a medical event and the medical language resources related to the related order information. An information processing device as described in any one of the appendices A1 to A4.

[0137] (Note A6) The extraction means is A second machine learning model, which has been trained to take the aforementioned medical consultation article as input and output words related to medical events contained in the aforementioned medical consultation article, A third machine learning model, trained to take the text related to the medical event and the related order information as input, and to output an embedding vector for the text and an embedding vector for a medical language resource related to the related order information, Using this method, the similarity between the embedding vector of the aforementioned text and the embedding vector of the medical language resource is referenced, and the event-related information is extracted. The information processing device described in Appendix A5.

[0138] (Note A7) The extraction means sorts the words related to the medical event in descending order of similarity and then extracts the event-related information. The information processing device described in Appendix A6.

[0139] (Note A8) The system further includes an output means for outputting the aforementioned event-related information. An information processing device as described in any one of the appendices A1 to A7.

[0140] [Additional Notes B] This disclosure includes the technologies described in the following appendices. However, the present invention is not limited to the technologies described in the following appendices, and various modifications are possible within the scope of the claims.

[0141] (Note B1) At least one processor refers to a set of order information containing multiple order information indicating each of multiple medical instructions, and performs a determination process to determine whether the set of order information corresponds to one or more types for which a detailed symptom description needs to be generated. In the determination process, if it is determined that the order information group corresponds to one or more types for which a detailed description of symptoms is required, at least one processor performs a identification process to identify related order information from the order information group that is associated with the corresponding type, The at least one processor performs an extraction process to extract from the medical examination article information related to the relevant order information, which is event-related information related to a medical event. The at least one processor performs a generation process to generate a symptom description by inputting the relevant order information and the event-related information into a language model, Information processing methods including

[0142] (Note B2) In the determination process, the at least one processor takes the order information set as input and uses a first machine learning model trained to output a determination result that determines whether the order information set corresponds to one or more types for which a detailed symptom description needs to be generated, to determine whether the order information set corresponds to one or more types for which a detailed symptom description needs to be generated. The information processing method described in Appendix B1.

[0143] (Note B3) The first machine learning model is further trained to take the order information set as input and output type information indicating the type to which the order information set belongs, if it determines that the order information set belongs to one or more types to which a detailed description of symptoms is required. The information processing method described in Appendix B2.

[0144] (Note B4) The first machine learning model is further trained to take the order information set and the type information as input and output related order information from the order information set that is related to the type indicated by the type information, In the specified process, the at least one processor identifies the related order information by inputting the order information set and the type information into the first machine learning model. The information processing method described in Appendix B3.

[0145] (Note B5) In the extraction process, the at least one processor references the similarity between the wording included in the medical examination article that relates to a medical event and the medical language resources related to the related order information, and extracts the event-related information. The information processing method described in any one of the appendices B1 to B4.

[0146] (Note B6) In the extraction process, the at least one processor, A second machine learning model, which has been trained to take the aforementioned medical consultation article as input and output words related to medical events contained in the aforementioned medical consultation article, A third machine learning model, trained to take the text related to the medical event and the related order information as input, and to output an embedding vector for the text and an embedding vector for a medical language resource related to the related order information, Using this method, the similarity between the embedding vector of the aforementioned text and the embedding vector of the medical language resource is referenced, and the event-related information is extracted. The information processing method described in Appendix B5.

[0147] (Note B7) In the extraction process, the at least one processor sorts the words related to the medical event in descending order of similarity and then extracts the event-related information. The information processing method described in Appendix B6.

[0148] (Note B8) The at least one processor further includes output processing that outputs the event-related information, The information processing method described in any one of the appendices B1 through B7.

[0149] [Additional Note C] This disclosure includes the technologies described in the following appendices. However, the present invention is not limited to the technologies described in the following appendices, and various modifications are possible within the scope of the claims.

[0150] (Note C1) A program that makes a computer function as an information processing device. The aforementioned computer, A determination means that refers to a group of order information containing multiple order information indicating each of multiple medical instructions, and determines whether the group of order information corresponds to one or more types for which a detailed description of symptoms needs to be generated, If the determination means determines that the order information group corresponds to one or more types for which a detailed description of symptoms is required, the determination means identifies related order information from the order information group that is associated with the corresponding type, An extraction means for extracting event-related information related to the medical event, which is related to the aforementioned related order information, from the medical examination article, A generation means for generating a detailed description of symptoms by inputting the aforementioned related order information and the aforementioned event-related information into a language model, An information processing program that functions as such.

[0151] (Note C2) The determination means takes the order information set as input and uses a first machine learning model trained to output a determination result that determines whether the order information set corresponds to one or more types for which a detailed symptom description needs to be generated, to determine whether the order information set corresponds to one or more types for which a detailed symptom description needs to be generated. The information processing program described in Appendix C1.

[0152] (Note C3) The first machine learning model is further trained to take the order information set as input and output type information indicating the type to which the order information set belongs, if it determines that the order information set belongs to one or more types to which a detailed description of symptoms is required. The information processing program described in Appendix C2.

[0153] (Note C4) The first machine learning model is further trained to take the order information set and the type information as input and output related order information from the order information set that is related to the type indicated by the type information, The identification means identifies the related order information by inputting the order information set and the type information into the first machine learning model. The information processing program described in Appendix C3.

[0154] (Note C5) The extraction means extracts event-related information by referring to the similarity between the wording contained in the medical examination article that relates to a medical event and the medical language resources related to the related order information. An information processing program described in any one of the appendices C1 to C4.

[0155] (Appendix C6) The extraction means is A second machine learning model, which has been trained to take the aforementioned medical consultation article as input and output words related to medical events contained in the aforementioned medical consultation article, A third machine learning model, trained to take the text related to the medical event and the related order information as input, and to output an embedding vector for the text and an embedding vector for a medical language resource related to the related order information, Using this method, the similarity between the embedding vector of the aforementioned text and the embedding vector of the medical language resource is referenced, and the event-related information is extracted. The information processing program described in Appendix C5.

[0156] (Note C7) The extraction means sorts the words related to the medical event in descending order of similarity and then extracts the event-related information. The information processing program described in Appendix C6.

[0157] (Note C8) The aforementioned computer, The aforementioned event-related information is further configured to function as an output means for outputting event-related information. An information processing program described in any one of the appendices C1 through C7.

[0158] [Additional Note D] This disclosure includes the technologies described in the following appendices. However, the present invention is not limited to the technologies described in the following appendices, and various modifications are possible within the scope of the claims.

[0159] (Note D1) It comprises at least one processor, and the at least one processor is A determination process that refers to a group of order information containing multiple order information indicating each of multiple medical instructions, and determines whether the group of order information corresponds to one or more types for which a detailed description of symptoms needs to be generated, In the determination process, if it is determined that the order information group corresponds to one or more types for which a detailed description of symptoms is required, a determination process is performed to identify related order information from the order information group that is associated with the corresponding type. An extraction process to extract event-related information related to the medical event, which is related to the aforementioned related order information, from the medical examination report, A generation process that generates a detailed description of symptoms by inputting the aforementioned related order information and the aforementioned event-related information into a language model, An information processing device that performs the following actions.

[0160] The information processing device may also include memory. Furthermore, the memory may store a program that causes at least one processor to execute each of the aforementioned processes.

[0161] (Note D2) In the determination process, the at least one processor takes the order information set as input and uses a first machine learning model trained to output a determination result that determines whether the order information set corresponds to one or more types for which a detailed symptom description needs to be generated, to determine whether the order information set corresponds to one or more types for which a detailed symptom description needs to be generated. The information processing device described in Appendix D1.

[0162] (Note D3) The first machine learning model is further trained to take the order information set as input and output type information indicating the type to which the order information set belongs, if it determines that the order information set belongs to one or more types to which a detailed description of symptoms is required. The information processing device described in Appendix D2.

[0163] (Note D4) The first machine learning model is further trained to take the order information set and the type information as input and output related order information from the order information set that is related to the type indicated by the type information, In the specified process, the at least one processor identifies the related order information by inputting the order information set and the type information into the first machine learning model. The information processing device described in Appendix D3.

[0164] (Note D5) In the extraction process, the at least one processor references the similarity between the wording included in the medical examination article that relates to a medical event and the medical language resources related to the related order information, and extracts the event-related information. An information processing device as described in any one of the appendices D1 to D4.

[0165] (Note D6) In the extraction process, the at least one processor, A second machine learning model, which has been trained to take the aforementioned medical consultation article as input and output words related to medical events contained in the aforementioned medical consultation article, A third machine learning model, trained to take the text related to the medical event and the related order information as input, and to output an embedding vector for the text and an embedding vector for a medical language resource related to the related order information, Using this method, the similarity between the embedding vector of the aforementioned text and the embedding vector of the medical language resource is referenced, and the event-related information is extracted. The information processing device described in Appendix D5.

[0166] (Note D7) In the extraction process, the at least one processor sorts the words related to the medical event in descending order of similarity and then extracts the event-related information. The information processing device described in Appendix D6.

[0167] (Note D8) The aforementioned at least one processor, Further output processing is performed to output the aforementioned event-related information. An information processing device as described in any one of the appendices D1 to D7.

[0168] [Additional Note E] This disclosure includes the technologies described in the following appendices. However, the present invention is not limited to the technologies described in the following appendices, and various modifications are possible within the scope of the claims.

[0169] (Note E1) A program that makes a computer function as an information processing device. To the aforementioned computer, A determination process that refers to a group of order information containing multiple order information indicating each of multiple medical instructions, and determines whether the group of order information corresponds to one or more types for which a detailed description of symptoms needs to be generated, If the determination process determines that the order information group corresponds to one or more types for which a detailed symptom description needs to be generated, then the identification process identifies related order information from the order information group that is associated with the corresponding type. An extraction process to extract event-related information related to the medical event, which is related to the aforementioned related order information, from the medical examination report, A generation process that generates a detailed description of symptoms by inputting the aforementioned related order information and the aforementioned event-related information into a language model, A non-temporary recording medium that stores an information processing program that executes that program. [Explanation of Symbols]

[0170] 1. 1A Information Processing Device 11 Judgment section 12 Specific part 13 Extraction part 14 Generation part 15 Output section 16 Acquisition Department ERI Event-Related Information LM Language Model MEI Medical Examination Information OIG Order Information Group OI Order Information Detailed description of SD symptoms TM1 First Machine Learning Model TM2, the second machine learning model TM3: The third machine learning model TY type information

Claims

1. A determination means that refers to a group of order information containing multiple order information indicating each of multiple medical instructions, and determines whether the group of order information corresponds to one or more types for which a detailed description of symptoms needs to be generated, If the determination means determines that the order information group corresponds to one or more types for which a detailed description of symptoms is required, the determination means identifies related order information from the order information group that is associated with the corresponding type, An extraction means for extracting event-related information related to the medical event, which is related to the aforementioned related order information, from the medical examination article, A generation means for generating a detailed description of symptoms by inputting the aforementioned related order information and the aforementioned event-related information into a language model, An information processing device equipped with the following features.

2. The determination means takes the order information set as input and uses a first machine learning model trained to output a determination result that determines whether the order information set corresponds to one or more types for which a detailed symptom description needs to be generated, to determine whether the order information set corresponds to one or more types for which a detailed symptom description needs to be generated. The information processing apparatus according to claim 1.

3. The first machine learning model is further trained to take the order information set as input and output type information indicating the type to which the order information set belongs, if it determines that the order information set belongs to one or more types to which the generation of symptom details is required. The information processing apparatus according to claim 2.

4. The first machine learning model is further trained to take the order information set and the type information as input and output related order information from the order information set that is related to the type indicated by the type information, The identification means identifies the related order information by inputting the order information group and the type information into the first machine learning model. The information processing apparatus according to claim 3.

5. The extraction means extracts event-related information by referring to the similarity between the wording contained in the medical examination article that relates to a medical event and the medical language resources related to the related order information. The information processing apparatus according to any one of claims 1 to 4.

6. The extraction means is A second machine learning model, which has been trained to take the aforementioned medical consultation article as input and output words related to medical events contained in the aforementioned medical consultation article, A third machine learning model, trained to take words related to the medical event and related order information as input, and to output an embedding vector for the words and an embedding vector for medical language resources related to the related order information, Using this method, the similarity between the embedding vector of the aforementioned text and the embedding vector of the medical language resource is referenced, and the event-related information is extracted. The information processing apparatus according to claim 5.

7. The extraction means sorts the words related to the medical event in descending order of similarity and then extracts the event-related information. The information processing apparatus according to claim 6.

8. The system further includes an output means for outputting the aforementioned event-related information. The information processing apparatus according to any one of claims 1 to 4.

9. At least one processor refers to a set of order information containing multiple order information indicating each of multiple medical instructions, and performs a determination process to determine whether the set of order information corresponds to one or more types for which a detailed symptom description needs to be generated. In the determination process, if it is determined that the order information group corresponds to one or more types for which a detailed description of symptoms is required, at least one processor performs a identification process to identify related order information from the order information group that is associated with the corresponding type, The at least one processor performs an extraction process to extract from the medical examination article information relating to the relevant order information, which is event-related information relating to a medical event. The at least one processor performs a generation process to generate a symptom description by inputting the relevant order information and the event-related information into a language model, Information processing methods including

10. A program that makes a computer function as an information processing device. The aforementioned computer, A determination means that refers to a group of order information containing multiple order information indicating each of multiple medical instructions, and determines whether the group of order information corresponds to one or more types for which a detailed description of symptoms needs to be generated, If the determination means determines that the order information group corresponds to one or more types for which a detailed description of symptoms is required, the determination means identifies related order information from the order information group that is associated with the corresponding type, An extraction means for extracting event-related information related to the medical event, which is related to the aforementioned related order information, from the medical examination article, A generation means for generating a detailed description of symptoms by inputting the aforementioned related order information and the aforementioned event-related information into a language model, An information processing program that functions as such.

Images