Medical care assistance system, display device, and medical care assistance program

Abstract

To provide a medical care assistance system, a display device, and a medical care assistance program with which it is possible to create a high quality medical care record based on dialog information between a veterinarian and an owner in veterinary medicine and increase the efficiency of veterinary medicine.SOLUTION: A medical care assistance system 1 converts voice information of conversion between a veterinarian and a pet owner into text data by a text data conversion unit 106. The text data and a template containing indications to a large-scale language model 205 are inputted as parameters to a medical care assistance information creation unit 204, and medical care assistance information is automatically created from the text data on the basis of the large-scale language model 205. Meanwhile, the created medical care assistance information is shown on a monitor 3 by a display device 30.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a medical care support system, a display device, and a medical care support program, and more particularly to a medical care support system, a display device, and a medical care support program that can create high-quality medical records based on conversation information between veterinarians and owners in animal medical care, thereby improving the efficiency of animal medical care. [Background technology]

[0002] Pets, including dogs and cats, are irreplaceable companions to humans, bringing joy to their owners' daily lives and sometimes even providing emotional support. In recent years, owners have become increasingly concerned about their pets' health, leading to an expanding demand for high-quality veterinary care.

[0003] However, because veterinary care is outside the scope of the national health insurance system, the cost of treating pets for injuries and illnesses tends to be higher than that of human care. Furthermore, while veterinary care uses diagnostic and treatment methods similar to those used in human care, pets cannot directly communicate their condition or treatment requests to veterinarians, so active testing is required to identify pathologies. Furthermore, owners must remember to follow the treatment plan at home and continue administering medication and properly monitor progress, meaning the owner's understanding of the treatment can easily affect the quality of veterinary care.

[0004] Therefore, in the field of animal medical care, so-called informed consent is important, where veterinarians provide owners with a thorough explanation of proposals and confirmations regarding treatment and examinations, as well as treatment costs, etc., and the owner agrees to the treatment method etc. after fully understanding the explanation. This consensus-building is considered important in improving the quality of animal medical care.

[0005] In order to reach an agreement between veterinarians and owners, a common method is to create medical records, including a chart that compiles information about the pet's treatment, and record the details of the treatment. In particular, in recent years, the introduction of electronic medical records has been promoted in veterinary medicine, just as in human medicine, in order to improve the efficiency of medical work and the accuracy of treatment through information sharing.

[0006] Meanwhile, in animal medical settings, during consultations, a nurse typically holds the patient's animal on the examination table while the veterinarian interviews the owner and explains the treatment plan. Therefore, creating medical records at the same time as the consultation is difficult due to personnel and physical constraints. In particular, when there are other pet owners waiting to be seen, sales efforts are required to shorten the waiting time for the customer, which adds to the time pressure, making it even more difficult to create medical records of a consistent quality.

[0007] To address this issue, for example, Patent Document 1 discloses a conversation recording system that stores audio information of conversations between medical professionals and patients, converts the audio information into text data for each speaker, and stores the converted text data in a dialogue format by the speakers as text data. This allows medical professionals to focus on taking medical interviews with patients and explaining treatment plans, and after explaining the treatment to the patient, they can create patient charts while checking the text data, making it possible to create high-quality medical records. [Prior art documents] [Patent documents]

[0008] [Patent Document 1] Japanese Patent Application Publication No. 2018-206055 Summary of the Invention [Problem to be solved by the invention]

[0009] However, the conversation recording system disclosed in the aforementioned Patent Document 1 simply stores the speaker's voice information as text data, so the medical records must be created by the veterinarian themselves. Therefore, even with the conversation recording system of Patent Document 1, medical professionals, including veterinarians, are still overwhelmed with creating documents such as medical records, and are unable to escape from the problem of long working hours.

[0010] In addition, veterinarians prepare medical records at their own discretion within a limited time frame, and there is little progress in standardizing the information to be written within the industry. This leads to incomplete information and input errors, which leads to a lack of information sharing during team medical care, which can lead to medical accidents.

[0011] Furthermore, explanations from veterinarians are not always easy for pet owners to understand, and in order to prevent problems with owners over treatment plans and costs, one of the medical services that is required is the creation of summaries for pet owners that summarize medical records in an easy-to-understand manner.

[0012] The present invention was devised in consideration of the above points, and aims to provide a medical support system, display device, and medical support program that can create high-quality medical records based on conversation information between veterinarians and animal owners in animal medical care, thereby improving the efficiency of animal medical care. [Means for solving the problem]

[0013] In order to achieve the above-mentioned object, the medical support system of the present invention comprises an audio information storage unit that stores audio information of conversations between medical professionals and subjects, including subjects who are the target of medical treatment; a text data conversion unit that converts the audio information into text data; a template storage unit that stores predetermined templates containing instructions for a large-scale language model for creating medical support information including at least one of a medical record for medical professionals, a medical summary for the subject corresponding to the medical record, and a medical statement based on the medical treatment content; and a medical support information creation unit that creates the medical support information based on the large-scale language model using the text data and the template as input parameters.

[0014] By providing a voice information storage unit that stores voice information of conversations between medical professionals and subjects, including subjects who are the subject of medical treatment, the stored voice information can be used as data for subsequent calculation processing. Furthermore, by saving the voice information as evidence, it is possible to avoid disputes with subjects regarding treatment plans and treatment costs.

[0015] Furthermore, by providing a text data conversion unit that converts voice information into text data, voice information can be converted into text data.

[0016] In addition, by providing a template storage unit that stores predetermined templates containing instructions for a large-scale language model for creating medical support information including at least one of medical records for medical professionals, a summary for the subject corresponding to the medical records, and medical details based on the medical treatment content, it is possible to prevent the instructions for the large-scale language model from differing depending on the user (here, medical professionals), and to maintain a consistent quality of the output of the large-scale language model.

[0017] Furthermore, by providing a medical support information creation unit that creates medical support information based on a large-scale language model using text data and templates as input parameters, it is possible to automatically create input information for each input item of the medical support information from the content of conversations during medical treatment. Therefore, since the user does not need to create the medical support information themselves, they can concentrate on examining and treating the subject, and the efficiency of medical work can be improved.

[0018] In addition, the input items of the medical record in the medical support information include at least one of the following input items: chief complaint, diagnosis, treatment plan, treatment, prescription, and explanation of test results. When the medical support information creation unit creates input information for each input item of the medical record based on a large-scale language model, it can automatically create the medical record based on the large-scale language model in accordance with the input items of the medical record that have been prepared in advance.

[0019] In addition, the input items for the medical support information summary include at least one input item for the method of administering prescribed medication to the subject, dietary management of the subject, monitoring of the subject's condition, management of the subject's living environment, and the subject's next scheduled medical appointment date, and when the medical support information creation unit creates input information for each input item of the summary based on a large-scale language model, it can automatically create the summary based on the large-scale language model in accordance with the input items of the summary for the subject that have been prepared in advance.

[0020] The system also has a fee memory unit that associates medical fees with medical items, and the medical support information creation unit calls up the medical fees corresponding to the medical items extracted based on the large-scale language model from the fee memory unit, and when creating input information for the input fields of the medical statement, it can automatically create the medical statement in accordance with the medical items extracted from the text data by the large-scale language model.

[0021] Furthermore, if the text data conversion unit has a phrase extraction unit that extracts specified phrases based on predetermined criteria from text data obtained from audio information, a phrase determination unit that determines whether the phrase extracted by the phrase extraction unit is correct from the context of a sentence containing the phrase, and a phrase conversion unit that converts the phrase into a correct phrase based on the context of the sentence containing the phrase if the phrase determination unit determines that the phrase is incorrect, erroneous conversion when converting audio information into text data can be prevented, thereby making it possible to create higher quality medical support information.

[0022] The system also has a medical support information storage unit that stores the medical support information created by the medical support information creation unit, and when the next medical examination of the subject is performed, if basic information of the subject is obtained from the medical support information stored in the medical support information storage unit, the basic information including the subject's weight, body temperature, and any information on past medical treatment details, prescription details, etc. can be reflected in the medical support information created for the current examination. Therefore, when creating the current medical support information, there is no need to input the basic information of the subject, and medical support information reflecting the previous medical treatment details can be created, which reduces the calculation load when creating medical support information and makes it possible to create higher quality medical support information.

[0023] Furthermore, if the audio information stored in the audio information storage unit is subjected to a predetermined anti-tampering process, the evidential value of the audio information can be increased.

[0024] In order to achieve the above-mentioned object, the display device of the present invention comprises a first display unit that creates time-series dialogue-style text data for each speaker based on audio information of conversations between medical professionals and subjects, including subjects who are the target of medical treatment, and displays the text data; a second display unit that displays at least one of medical support information obtained from the text data based on a large-scale language model, the medical support information including a medical record for the medical professional, a medical summary for the subject corresponding to the medical record, and a medical statement based on the medical treatment content; a selection unit that selects one of the medical support information to be displayed on the second display unit; and a display control unit that displays the one medical support information selected by the selection unit on the second display unit.

[0025] Here, by creating dialogue-style text data for each speaker in chronological order based on audio information of conversations between medical professionals and subjects, including medical examinees, and providing a first display unit that displays the text data, the content of the conversation between the medical professional and the subject is displayed in chronological order, allowing the identification of speakers and the content of the conversation to be confirmed retroactively. Furthermore, any part of the conversation can be easily identified by keyword search of the text data.

[0026] In addition, by providing a second display unit that displays at least one of the medical support information obtained from text data based on a large-scale language model, including a medical record for medical professionals, a summary for the subject corresponding to the medical record, and a medical statement based on the medical treatment content, the user can check the contents of the medical support information created by the large-scale language model.

[0027] In addition, by providing a selection unit that selects one piece of medical support information to be displayed on the second display unit from the medical support information, and a display control unit that displays the one piece of medical support information selected by the selection unit on the second display unit, any medical support information selected by the user can be displayed on the second display unit.

[0028] In addition, if the medical record displayed on the second display unit includes at least one of the following items created based on a large-scale language model: chief complaint, diagnosis, treatment plan, treatment, prescription, and explanation of test results, the user can confirm the medical details based on the contents of the medical record displayed on the second display unit.

[0029] Furthermore, if the summary displayed on the second display unit includes at least one of the following items created based on a large-scale language model: how to administer prescribed medication to the subject; dietary management of the subject; monitoring of the subject's condition; management of the subject's living environment; and the subject's next scheduled medical appointment date, the user can provide the subject with a detailed and easy-to-understand explanation of the medical treatment based on the summary displayed on the second display unit.

[0030] Furthermore, if the medical details displayed on the second display unit are medical fees corresponding to medical items extracted based on a large-scale language model, the user can provide the subject with a detailed and easy-to-understand explanation of the basis for treatment costs, etc., based on the medical details displayed on the second display unit.

[0031] In order to achieve the above-mentioned object, the medical support program of the present invention comprises the steps of acquiring audio information of a conversation between a medical professional and a subject, including a subject who is the target of medical treatment; converting the audio information into text data; selecting a template including instructions for a large-scale language model to create medical support information including at least one of a medical record for the medical professional, a medical summary for the subject corresponding to the medical record, and a medical statement based on the medical treatment content; inputting the text data and the template into the large-scale language model as input parameters; and creating the medical support information from the text data based on the large-scale language model to which the input parameters have been input.

[0032] Here, by providing a step of acquiring voice information of conversations between medical professionals and subjects, including subjects to be treated, the acquired voice information can be used as data for subsequent calculation processing. Furthermore, by storing the voice information as evidence, it is possible to avoid disputes with subjects regarding treatment plans and treatment costs.

[0033] Furthermore, by providing a step of converting voice information into text data, voice information can be converted into text data.

[0034] In addition, by providing a step of selecting a template including instructions for a large-scale language model for creating medical support information including at least one of a medical record for medical professionals, a summary for the subject corresponding to the medical record, and a medical statement based on the medical treatment content, it is possible to prevent the instructions for the large-scale language model from differing depending on the user, and to maintain a consistent quality of the output of the large-scale language model.

[0035] Furthermore, by providing a step of creating medical support information from text data based on a large-scale language model into which input parameters have been entered, the medical support information is automatically created from records of conversations during medical treatment, eliminating the need for the user to create the medical support information themselves, allowing the user to concentrate on examining and treating the subject and improving the efficiency of medical work. [Effects of the Invention]

[0036] The medical support system, display device, and medical support program of the present invention are capable of creating high-quality medical records based on conversational information between veterinarians and owners in animal medical care, thereby improving the efficiency of animal medical care. [Brief explanation of the drawings]

[0037] [Figure 1] 1 is a diagram showing a state in which a medical assistance system according to an embodiment of the present invention is connected to a terminal device via an Internet line. [Figure 2] FIG. 2 is a block diagram showing the internal configuration of the medical support system. [Figure 3] FIG. 2 is a block diagram showing the internal configuration of a text data conversion unit. [Figure 4] FIG. 10 is a diagram showing an example of an incorrect conversion list stored in a dictionary storage unit. [Figure 5] FIG. 10 is a diagram illustrating an example of data content stored in a template storage unit. [Figure 6] FIG. 4 is a diagram showing an example of data contents stored in a fee storage unit. [Figure 7] FIG. 2 is a block diagram showing the internal configuration of the display device. [Figure 8] FIG. 10 is a diagram showing an example of a display screen in which medical records, among the medical support information, are displayed on a monitor. [Figure 9] FIG. 10 is a diagram showing an example of a display screen in which a medical summary, out of medical support information, is displayed on a monitor. [Figure 10] FIG. 10 is a diagram showing an example of a display screen in which medical details, among medical support information, are displayed on a monitor. [Figure 11] FIG. 2 is a diagram showing a calculation flow of a medical assistance program according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0038] Hereinafter, embodiments of the present invention relating to a medical assistance system, a display device, and a medical assistance program will be described with reference to the drawings to facilitate understanding of the present invention.

[0039] First, an overview of the overall network configuration including a medical assistance system 1 according to an embodiment of the present invention will be described with reference to Figure 1. The medical assistance system 1 according to an embodiment of the present invention is a computer for running a medical assistance program, and can be, for example, a server installed in a facility that provides medical services (in an embodiment of the present invention, a veterinary clinic (hereinafter referred to as a "hospital") that provides medical services to subjects (hereinafter referred to as "pets") who are the subject of medical treatment), or a rental server (cloud server) on which the medical assistance program is installed.

[0040] The terminal device 2 is a personal computer used by a medical professional (a "veterinarian" in this embodiment of the present invention) during medical treatment, and the terminal device 2 and the medical support system 1 are connected to each other so that they can communicate with each other via an internet line 4. If the medical support system 1 is a cloud server or an on-premise server, multiple terminal devices 2 can simultaneously access the medical support system 1 via the internet line 4.

[0041] Here, the medical assistance system 1 and the terminal device 2 do not necessarily need to be connected to each other via the Internet line 4, and may be connected by any communication means. Furthermore, the terminal device 2 is not particularly limited as long as it is an information terminal having a communication function, such as a personal computer, a mobile terminal, a tablet terminal, or the like.

[0042] Each component of the medical assistance system 1 will be described in detail below with reference to Fig. 2. The medical assistance system 1 is composed of a voice information system unit 10 that records the content of conversations mainly between veterinarians and subjects (pet owners in the embodiment of the present invention) as voice information and converts it into text data, a medical assistance information system unit 20 that creates medical assistance information such as medical records based on the text data processed by the voice information system unit 10, and a display device 30 that displays the text data of the conversations and the medical assistance information on the monitor 3 of the terminal device 2.

[0043] 1. Voice Information Systems Department First, the main configuration of the voice information system unit 10 will be described below.

[0044] [Audio input section] The voice input unit 101 has a function of collecting the conversation between the veterinarian and the owner during medical treatment using a microphone (not shown), accepting the input of the collected voice, and acquiring voice information. The voice information acquired by the voice input unit 101 is stored in the voice information storage unit 103 via the voice information transmission / reception unit 102.

[0045] [Audio input control section] The voice input control unit 104 has a function of controlling the start and pause of voice input to the voice input unit 101, and also of analyzing voice information input to the voice input unit 101 and performing predetermined control.

[0046] For example, when the voice input control unit 104 detects that the veterinarian has uttered the word "start," it instructs the voice input unit 101 to start acquiring voice information. Also, when it detects that the veterinarian has uttered the word "stop," it instructs the voice input unit 101 to stop acquiring voice information. In this way, it is possible to control the start and stop of acquiring voice information based on voice instructions from the veterinarian without operating the terminal device 2.

[0047] Here, the words used to control the start and stop of voice information acquisition are not limited to "start" and "stop", but specific words may be registered in advance, and the start or stop of voice information acquisition may be controlled based on the registered language.

[0048] It is also possible to control the start and stop of acquisition of audio information by operating the "start" button or "stop" button displayed on the monitor 3 of the terminal device 2. However, by issuing instructions to the audio input unit 101 based on audio instructions from the veterinarian, there is no need to stop and interrupt the examination to operate the equipment, which can improve the efficiency of medical treatment work.

[0049] Furthermore, the voice input control unit 104 analyzes the voice information received by the voice input unit 101, identifies the language used in the conversation and the number of main speakers in the conversation, and if it determines that the volume of the voice information is low or that the voice information has been interrupted due to a deterioration in the communication environment, it generates an error code and notifies the veterinarian that there is a problem with the voice information and that the voice input has not been made correctly. The notification means here is not particularly limited, but may, for example, display a warning screen to that effect on the monitor 3 of the terminal device 2, or sound a warning.

[0050] Furthermore, when the voice input control unit 104 detects a silence period during which no voice input is detected for a certain period of time (for example, 2 to 3 seconds) in the voice input unit 101, it recognizes the conversation before and after the silence period as one phrase. Note that the detection of phrases may be performed based on real-time voice information input to the voice input unit 101, or may be performed after acquiring voice information of the entire conversation from the start to the end.

[0051] [Tamper-proof section] The tampering prevention unit 105 has a function of performing tampering prevention processing to prevent tampering of the audio information stored in the audio information storage unit 103. In the tampering prevention processing, for example, when audio information is stored in the audio information storage unit 103, a hash code is first generated from the content using a cryptographic hash function. Next, a digital signature is applied to this hash code using the creator's private key, and a timestamp is added. This makes it possible to effectively detect tampering of the audio file. This hash code is generated using a cryptographic hash function, and the hash function used in this embodiment of the present invention is SHA-256.

[0052] In this embodiment of the present invention, the hash function used is SHA-256 due to its strong security characteristics and proven adoption in a wide range of fields, but the hash function is not necessarily limited to this and can be selected from any commonly used hash function. For example, SHA-3 or other hash functions conforming to industry standards can also be applied.

[0053] Here, it is not necessary to apply tamper-proof processing to the audio information stored in the audio information storage unit 103. However, by applying tamper-proof processing, the evidentiary value of the audio information can be increased, so that when a dispute arises between a veterinarian and an owner over treatment policy or money, for example, the audio information can be used as circumstantial evidence to resolve the dispute before it occurs.

[0054] [Text data conversion section] The text data conversion unit 106 has the function of transcribing the voice information input from the voice input unit 101 using artificial intelligence that uses a trained machine learning model, and then converting it into text data.

[0055] Figure 3 shows a block diagram of the internal configuration of the text data conversion unit 106. As shown in Figure 3, the text data conversion unit 106 is made up of a machine learning model 120 and an erroneous conversion prevention unit 130 that prevents erroneous conversions such as homonyms during transcription. The processing results by the erroneous conversion prevention unit 130 are fed back to the machine learning model 120 for additional learning, and the algorithm of the machine learning model 120 is optimized.

[0056] The misconversion prevention unit 130 is mainly composed of a word extraction unit 131, a word determination unit 132, and a word conversion unit 133, and further has a dictionary storage unit 134 in which a list of misconversions that are likely to occur in medical terminology (see Figure 4) is stored in advance.

[0057] For example, when it is detected that the speech information contains the word "tumor" registered in the incorrect conversion list, the word extraction unit 131 extracts the word "tumor." Then, the phrase determination unit 132 determines whether "tumor" should be converted to "main" in conversion 1 or "tumor" in conversion 2, based on the context before and after the sentence containing the word "tumor." If the determination result shows that "tumor" in conversion 2 is correct, the phrase conversion unit 133 converts the speech information of "tumor" into text data of "tumor."

[0058] Furthermore, by pre-registering in the dictionary storage unit 134 not only a list of mistranslations consisting of homonyms, but also difficult drug names, medical terms, or words and phrases that are difficult to distinguish by ear, it is possible to prevent mistranslations of words and phrases that are difficult to distinguish from audio information.

[0059] For example, there may be a case where a veterinarian pronounces the name of a drug as "doxycycline," but the initial "do" is recognized as a "to" in the audio information, resulting in the erroneous conversion to text data as "toxycycline." Even in this case, the phrase extraction unit 131 extracts the word "toxycycline," and the phrase determination unit 132 determines from the context that the veterinarian is explaining a drug, and then determines that "toxycycline" should be converted to "doxycycline," and the phrase conversion unit 133 converts it to text data.

[0060] If it is necessary to add homonyms or technical terms that are not in the incorrect conversion list, the administrator of the medical support system 1 (a veterinarian in the embodiment of the present invention) can add words or phrases, and by having the machine learning model 120 learn each time, it is possible to improve the accuracy of text data conversion.

[0061] As described above, the text data converted from the voice information is transmitted from the text data transmitting / receiving unit 107 to the text data storage unit 108 and stored therein. The text data stored in the text data storage unit 108 is then displayed on the monitor 3 of the terminal device 2 through the display device 30, which will be described later.

[0062] 2. Medical Support Information Systems Department The medical support information system unit 20 creates medical support information, including medical records, based on the text data processed by the voice information system unit 10. The main components of the medical support information system unit 20 will be described below.

[0063] [Template storage section] The template storage unit 201 stores instructions to the large-scale language model 205, which is an artificial intelligence for creating medical support information, as templates in advance. The medical support information includes medical records that function as electronic medical records, a summary that briefly summarizes the medical treatment details to be presented to the owner, and various documents such as medical bills calculated based on the medical treatment details (see Figs. 8 to 10 for examples of medical support information).

[0064] Here, the medical record items include "Chief complaint," "Owner's anxiety," "Differentiation of similar diseases, suspected injury or illness, diagnosis," "Examination direction, test name, test results, treatment details," "Treatment policy, treatment plan," "Prescription," "Next visit," and "Other (e.g., explanation of test results)."

[0065] The summary also includes items such as "compliance with medication," "dietary management," "regular health checks," "seizure monitoring," and "environmental management."

[0066] The medical bill includes items such as "product name," "product code," "quantity," and "standard price."

[0067] Here, the medical support information is not necessarily limited to the medical record, abstract, and medical statement described above, but may be composed of any one or a combination of two or more of the medical record, abstract, and medical statement, or documents containing other information related to medical treatment may be added as medical support information as appropriate.

[0068] Fig. 5 shows an example of data contents stored in the template storage unit 201. The template storage unit 201 is composed of a main table and a sub-table, and as shown in Fig. 5(a), the main table manages the template name and the category of medical support information using individual IDs. Also, as shown in Fig. 5(b), the sub-table manages specific instructions (prompts) for input items of medical support information and the ID of the parent table linked to them using individual IDs.

[0069] The prompts registered in the sub-table are specific instructions to the large-scale language model 205 for each input item, such as "Please summarize the conversation focusing on <the owner's chief complaint>. Please write the purpose of today's consultation from the owner's perspective" for the input item "Chief complaint" of medical information. Other instructions include, for example, "Do not use honorifics," "Do not make assumptions, answer only facts that are clear from the content of the conversation," and "Record the name of the pet, and write people's names as <owner> or <veterinarian>." Templates containing various prompts for creating medical support information are registered.

[0070] Here, in order to improve the accuracy of the output by the large-scale language model 205, it is preferable to register more specific instruction content as in the above-mentioned template, but Figure 5 is just an example and is not necessarily limited to the content of such a template. The content of the template registered in the sub-table can be modified as appropriate by the veterinarian depending on the type and content of the medical care support information to be created.

[0071] Furthermore, the data structure of the template storage unit 201 does not necessarily have to be composed of a main table and sub-tables. For example, the main table and sub-tables may be integrated and stored as a single table data, or may be stored in a data format other than table data.

[0072] [Clinical Analysis Department] The medical care analysis unit 202 has a function of identifying a template that is deemed appropriate for the pet's medical care based on the conversation content converted into text data, and calling up the identified template from the template storage unit 201.

[0073] It should be noted that the template does not necessarily need to be called from the template storage unit 201 via the medical analysis unit 202, and any template selected by a veterinarian may be called, for example.

[0074] [Medical Information Acquisition Department] The medical information acquisition unit 203 acquires various data, including text data created by the voice information system unit 10, templates called up by the medical analysis unit 202, and, if the pet being treated is a follow-up visit, past medical record information stored in the medical support information storage unit 209 (described later).The acquired data is then organized as input parameters for the large-scale language model 205.

[0075] Here, various information such as basic information about the pet (name, weight, temperature, etc.), the date of the previous medical examination, the details of the previous diagnosis and treatment, etc. is acquired as medical record information and is reflected in the medical support information created for the current medical examination. Note that if the pet being treated is a first-time patient and has no previous medical history, the basic information about the pet may be entered manually by a veterinarian or by other predetermined input means.

[0076] Furthermore, it is not necessarily required to have the medical information acquisition unit 203, and the input parameters consisting of various data may be directly input to the medical support information creation unit 204 described below, organized as input parameters, and then output to the large-scale language model 205.

[0077] [Medical Support Information Creation Department] The medical support information creation unit 204 uses the various data acquired by the medical information acquisition unit 203 as input parameters and creates medical support information based on a large-scale language model 205, which is an artificial intelligence. The large-scale language model 205 is a language model that has been trained on a large amount of data, such as articles, books, and websites on the Internet. When the above-mentioned template is input, it has the function of analyzing the text data generated by the voice information system unit 10 and automatically generating input information for each input item of the medical support information.

[0078] [Artificial Intelligence Control Unit] The artificial intelligence control unit 206 has the function of tasking the process of sending a request to the large-scale language model 205 and managing the tasks until the process is completed so that the sent request is processed appropriately. For example, it plays a role in maintaining the quality of the generation process using the large-scale language model by notifying the predicted time required to create medical support information, managing the processing order of multiple requests, or managing how to deal with errors that occur in the input information being generated.

[0079] [Fee storage section] The fee storage unit 207 stores fees associated with medical treatment items, as shown in Fig. 6. The medical treatment items registered in the fee storage unit 207 can be added or deleted as needed by the veterinarian, and the fees for the medical treatment items can also be revised as needed in response to fee revisions.

[0080] [Fee Detail Analysis Section] The medical details analysis unit 208 has the function of extracting medical items contained in the text data as a result of analysis by the large-scale language model 205, and calculating the medical expenses for the extracted medical items by referring to the fee information stored in the fee memory unit 207, and creating a medical details.

[0081] Specifically, when the fee statement analysis unit 208 detects words related to medical treatment items such as "blood test," "abdominal ultrasound," and "urine test" in the text data, it references the fees corresponding to these words from the fee storage unit 207. Then, based on the reference result, it creates a medical statement as accounting information for this medical treatment.

[0082] Furthermore, as described above, the fee statement analysis unit 208 is not limited to extracting as medical statements only items for which the words in the text data exactly match the words stored in the fee storage unit 207, but can also extract medical items including words that partially match. For example, if the word "blood test" is detected in the text data, "blood test" and "blood biochemistry test" can also be extracted as candidate medical items as test items related to "blood." Note that if medical items are extracted including words that partially match, unnecessary medical items may also be reflected in the medical statement, but in that case, the veterinarian can delete the unnecessary medical items as appropriate.

[0083] As described above, the medical support information created by the large-scale language model 205 is stored in the medical support information storage unit 209. The medical support information storage unit 209 is composed of a medical record storage unit 209a, an abstract storage unit 209b, and a medical statement storage unit 209c, and the medical support information is stored in each storage unit.

[0084] 3.Display device The display device 30 has a function of displaying the medical support information created by the medical support information system unit 20 on the monitor 3 of the terminal device 2 based on an input signal from the medical support information system unit 20 .

[0085] 7 shows a block diagram of the display device 30. When an operation signal from a veterinarian is received by a display receiving unit 301, a display control unit 302 selects the display content to be displayed on the monitor 3 and transmits it to a display output unit 303. Then, based on the output signal from the display output unit 303, the display content is displayed on the monitor 3.

[0086] 8 to 10 are diagrams showing an example of a display screen in which medical care support information is displayed on the monitor 3. The monitor 3 has two main display sections, a first display section 31 and a second display section 32, around which basic information such as the veterinarian's name, the owner's name, and the pet's name is displayed, and operation buttons for performing predetermined operations are laid out.

[0087] First, the first display unit 31 displays text data showing the conversation content created by the voice information system unit 10. At this time, the text data showing the conversation content displayed on the first display unit 31 is displayed in chronological order, separated into phrases of the conversation for each speaker identified by the voice input control unit 104.

[0088] When the text data of each conversation displayed on the first display unit 31 is pressed with, for example, a mouse or a finger on the screen, the audio information of the pressed conversation is played. Also, when the "Play All Audio" button at the bottom of the first display unit 31 is pressed with the mouse or a finger on the screen, the audio information of the entire conversation is played. Therefore, the veterinarian can check the content of the text data along with the audio information.

[0089] Next, the medical support information created by the medical support information system unit 20 is displayed on the second display unit 32. A selection unit 33 for selecting medical record information to be displayed on the second display unit 32 is provided above the second display unit 32, and when the veterinarian presses one of the selection buttons for "medical record," "summary," or "medical details" displayed on the selection unit 33, the input signal is received by the display receiving unit 301, the selected medical record information is called from the medical support information storage unit 209, and is output from the display output unit 303 to the monitor 3.

[0090] Here, Figure 8 shows the screen when "Medical Record" is selected in the selection section 33, Figure 9 shows the screen when "Summary" is selected in the selection section 33, and Figure 10 shows the screen when "Medical Details" is selected in the selection section 33.

[0091] Of the medical support information, the medical records are used as electronic medical records that are checked by veterinarians and other medical professionals, and the summary is printed out as needed and provided to the owner at the end of the treatment. In addition, the medical statement is linked to the hospital's accounting system and used at the time of payment at the end of the treatment.

[0092] Furthermore, if the veterinarian checks the text data of the medical support information displayed on the monitor 3 and determines that further correction of mistranslations or expressions is necessary, he or she corrects the text data as appropriate based on the administrator authority of the terminal device 2, and the corrected medical support information is stored in the medical support information storage unit 209 either as an overwrite of the original data or as a separate file.

[0093] The above is the configuration of the medical assistance system 1 and the display device 30 according to the embodiment of the present invention. Next, the medical assistance program executed by the medical assistance system 1 will be described with reference to the flowchart of FIG.

[0094] First, the voice input unit 101 acquires voice information of a conversation between a veterinarian and an owner during medical treatment, and the acquired voice information is stored in the voice information storage unit 103 (STEP 1).

[0095] Next, the voice information stored in the voice information storage unit 103 is converted into text data by the text data conversion unit 106 (STEP 2).

[0096] Next, a template of instruction content to be created for the large-scale language model 205 is selected from the templates stored in the template storage unit 201 (STEP 3).

[0097] Next, the text data and the template are input as input parameters to the large-scale language model 205 of the medical support information creation unit 204 (STEP 4).

[0098] Then, based on the input values in STEP 4, medical support information is created from the text data based on the large-scale language model 205 (STEP 5).

[0099] By repeating the calculations in steps 1 to 5, medical care support information can be automatically created based on the audio information of the conversation between the veterinarian and the owner.

[0100] The above is a description of the medical assistance system, display device, and medical assistance program according to an embodiment of the present invention. While the above description has been given of an embodiment in which medical assistance information is created based on a conversation between a veterinarian and a pet owner at a veterinary clinic, the present invention can also be applied to other embodiments. For example, when the present invention is applied to human medicine, medical assistance information can be created based on a conversation between a doctor and a patient.

[0101] As described above, the medical support system, display device, and medical support program of the present invention are capable of creating high-quality medical records based on conversational information between veterinarians and owners in animal medical care, thereby improving the efficiency of animal medical care. [Explanation of symbols]

[0102] 1 Medical support system 10. Voice Information Systems Department 101 Audio input section 102 Voice information transmission and reception unit 103 Voice information storage unit 104 Audio input control unit 105 Tamper-proof section 106 Text data conversion unit 107 Text data transmission and reception unit 108 Text data storage unit 20 Medical Support Information Systems Department 201 Template memory section 202 Medical Analysis Department 203 Medical Information Acquisition Department 204 Medical Support Information Creation Department 205 Large-scale Language Models 206 Artificial Intelligence Control Unit 207 Fee Memory Unit 208 Medical details analysis department 209 Medical support information storage unit 30 Display device 301 Display receiver 302 Display control unit 303 Display output unit 2. Terminal Device 3 monitors 4. Internet connection

Claims

1. a voice information storage unit that stores voice information of conversations between medical professionals and subjects, including subjects who are medical treatment targets; a text data conversion unit that converts the voice information into text data; a template storage unit that stores predetermined templates including instructions for a large-scale language model for creating medical support information including at least one of a medical record for a medical professional, a medical summary for a subject corresponding to the medical record, and a medical statement based on the medical treatment content; a medical support information creation unit that creates the medical support information based on the large-scale language model using the text data and the template as input parameters. Medical support system.

2. The input items of the medical record of the medical support information include at least one input item of a chief complaint, a diagnosis, a treatment plan, a treatment, a prescription, and an explanation of test results, The medical support information creation unit creates input information for each input item of the medical record based on the large-scale language model. The medical support system according to claim 1 .

3. The input items of the summary of the medical support information include at least one input item of a method for administering prescribed medicine to the subject, dietary management of the subject, monitoring of the subject's condition, management of the subject's living environment, and the subject's next scheduled medical treatment date; The medical support information creation unit creates input information for each input item of the abstract based on the large-scale language model. The medical support system according to claim 1 or 2.

4. a fee storage unit that associates medical fees with medical items; The medical support information creation unit retrieves medical fees corresponding to the medical items extracted based on the large-scale language model from the fee storage unit, and creates input information for the input items of the medical statement. The medical support system according to claim 1 or 2.

5. The text data conversion unit a phrase extraction unit that extracts specified phrases based on predetermined criteria from text data obtained from the speech information; a phrase determination unit that determines whether the phrase extracted by the phrase extraction unit is correct or not based on the context of a sentence including the phrase; a phrase conversion unit that converts the phrase into a correct phrase based on the context of a sentence containing the phrase when the phrase determination unit determines that the phrase is incorrect. The medical support system according to claim 1 or 2.

6. a medical support information storage unit that stores the medical support information created by the medical support information creation unit, When the subject is examined next time, basic information of the subject is acquired from the medical support information stored in the medical support information storage unit. The medical support system according to claim 1 or 2.

7. The voice information stored in the voice information storage unit has undergone a predetermined tamper-proof process. The medical support system according to claim 1 or 2.

8. a first display unit that creates time-series dialogue-style text data for each speaker based on audio information of conversations between a medical professional and a subject, including a subject to be treated, and displays the text data; a second display unit that displays at least one of medical support information obtained from the text data based on a large-scale language model, the medical support information including a medical record for a medical professional, a medical summary for a subject corresponding to the medical record, and a medical statement based on the medical treatment content; a selection unit that selects one piece of medical support information to be displayed on the second display unit from the medical support information; a display control unit that displays the one piece of medical support information selected by the selection unit on the second display unit. Display device.

9. The medical record displayed on the second display unit includes at least one item of a chief complaint, a diagnosis, a treatment plan, a treatment, a prescription, and an explanation of test results, which are created based on the large-scale language model. The display device according to claim 8 .

10. The summary displayed on the second display unit includes at least one of the following items: a method for administering prescribed medicine to the subject, dietary management of the subject, monitoring of the subject's condition, management of the subject's living environment, and the subject's next scheduled medical appointment date, which are created based on the large-scale language model. The display device according to claim 8 or 9.

11. The medical details displayed on the second display unit are medical fees corresponding to the medical items extracted based on the large-scale language model. The display device according to claim 8 or 9.

12. acquiring audio information of a conversation between a medical professional and a subject, including a subject to be treated; converting the audio information into text data; selecting a template including instructions for a large-scale language model for generating medical support information including at least one of a medical record for a medical professional, a summary for a subject corresponding to the medical record, and a medical statement based on the medical care content; inputting the text data and the template as input parameters into the large-scale language model; and creating the medical support information from the text data based on the large-scale language model to which the input parameters have been input. Clinical support program.

Images