Record management apparatus and method

Abstract

To make a data input screen for one user specified by a user name according to a speech including the user name.SOLUTION: When a user name is spoken, a record management server creates an analysis request including text of the user name and transmits the analysis request to a morpheme analysis server (steps 71, 73). The morpheme analysis server performs morpheme analysis on the text and transmits a reading kana of the person name (user name) to the record management server as an analysis result (step 75). The record management server transmits a user ID corresponding to the received reading kana to the record management terminal (step 74). The record management terminal displays a nursing care content input screen for the user ID (step 72).SELECTED DRAWING: Figure 9A

Description

Technical Field

[0001] This invention relates to a recording management apparatus and method.

Background Art

[0002] Patent Document 1 discloses an apparatus that automatically inputs values into each item of a form on a care record input screen using text data obtained by voice recognition.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

[0004] In the apparatus described in Patent Document 1, a template for extracting numerical values and the like from voice-recognized text data is prepared in advance. If there is a portion in the text data that matches the template, it is extracted from the text data and input as the value of the corresponding item on the care record input screen. The template includes predetermined terms such as "body temperature" and "blood pressure", and the item is specified by recognizing this predetermined term.

[0005] If the item name is pronounced first and then the data content of that item is pronounced, it is possible to input the desired data content (a numerical value representing body temperature) into the data input field for the desired item (for example, the body temperature input field). However, when the recording form includes a plurality of items (data input fields), if the item name and data content are pronounced for all of them, it takes time for data recording and the difficulty of operation at the site increases.

Summary of the Invention

Problems to be Solved by the Invention

[0006] An object of this invention is to enable easy input of appropriate data content into the data input field for an appropriate item.

[0007] In particular, the object of the present invention is to enable appropriate data content to be input into appropriate items without particularly considering the order of speech or the pronunciation of item names when speaking.

[0008] The data input support device according to the present invention includes a form name of a record form having a plurality of record items (names) and data input fields for each record item, and a microphone that receives an input of speech including the record content to be input into the data input fields, speech recognition means for converting the speech including the form name and the record content received by the microphone into a character string, and a plurality of record items (names) included in the record form specified by the record form name included in the character string converted by the speech recognition means and the record content (which may be the character string itself) included in the character string, and a receiving means for receiving the result of a sorting process in which the record content included in the speech is sorted into record items according to the content thereof, which is executed according to a processing request sentence including a sorting instruction for associating the record content with the record items, in a generative AI (generative AI) to which a processing request sentence is given.

[0009] The data input support method according to the present invention receives, by a microphone, an input of speech including a form name of a record form having a plurality of record items and input fields for each record item, and the record content to be input into the input fields, converts, by speech recognition means, the speech including the record form name and the record content received by the microphone into a character string, and receives, by receiving means, the result of a sorting process in which the record content included in the speech is sorted into record items according to the content thereof, which is executed according to a processing request sentence including a sorting instruction for associating the record content with the record items, in a generative AI to which a processing request sentence is given, the generative AI being specified by the record form name included in the character string converted by the speech recognition means and including a plurality of record items included in the record form and the record content included in the character string.

[0010] According to the present invention, since the recorded content (data content) included in the speech is sorted (associated) into recording items corresponding to the content, it is possible to input the recording content (data content) suitable for the recording item into the data input field for each recording item provided in the recording form. Since it is not necessary to repeatedly select the recording item (or its data input field) using an input device (such as a keyboard) and input data into the data input field of the selected recording item, the efficiency of the clerical work of recording is improved.

[0011] That is, since the recording form includes a plurality of recording items and data input fields for each recording item, in order to input the appropriate recording content into the appropriate recording item, it is necessary to select a specific recording item from among the plurality of recording items and input the recording content (for example, a numerical value) into the data input field of the selected recording item. In the case of data input using an input device (mouse, keyboard, etc.), a specific recording item (or its data input field) is selected in advance, for example, by clicking, and the recording content (for example, a numerical value) is input in that state. The selection of a specific recording item (selection of the data input field) is performed with one click, but if the recording form includes many recording items (data input fields), the number of clicks will increase. In the case of voice input, if the recording item name is pronounced in advance and then the recording content is pronounced, it is possible to input the desired recording content into the data input field of the desired recording item. However, if the recording form includes many recording items (data input fields) and the recording item name has to be pronounced for all of them, it still takes time and the difficulty of the operation at the data input site increases.

[0012] According to the present invention, a generative AI is used to realize the association (sorting process) between the desired recording item and the recording content without necessarily pronouncing the recording item (name). The present invention operates as a machine that executes a process (sorting process) of associating the recording content with the recording item using the generative AI, and for this, for example, ChatGPT can be used.

[0013] A processing request sentence including a plurality of record items (names), record contents (numerical values etc. to be input into the data input fields for each record item), and a sorting processing instruction is sent to the generative AI.

[0014] According to this invention, when a recording form name is received by a microphone, a processing request sentence including a plurality of record items (names) provided in the recording form specified by the recording form name is given to the generative AI. Also, since the record items (names) (candidates thereof) for which the record contents to be sorted in the processing request sentence should be accurately conveyed to the generative AI, the accuracy of the association (sorting) between the record items and the record contents can be improved.

[0015] When actually having ChatGPT perform the sorting process, it was confirmed that even if only the record contents are included during the conversation and the corresponding record item (name) is not included during the conversation, the desired record contents can be associated (sorted) with the desired record items with relatively high accuracy. The sorting process is accurately executed even for spoken language, and the clerical efficiency of data entry is greatly improved. Although ChatGPT has been proposed for use in tasks such as proofreading and correcting articles, summarizing articles and concepts, brainstorming, researching and identifying arguments, proposing ideas, programming, English to Japanese translation, Japanese to English translation, writing compositions, etc., it has not been confirmed as a fact that ChatGPT is currently proposed for use in the sorting process of associating appropriate record contents with appropriate record items.

[0016] The speech recognition process for converting speech into a character string is executed on a recording management terminal equipped with a microphone. For example, a virtual assistant (such as Siri, Google Now, etc.) included in the operating system of the recording management terminal can be used for character string conversion (speech-to-text).

[0017] The above-mentioned record form preferably includes a form for nursing care records or support records for the elderly or disabled, and a form for personal records of the elderly or disabled themselves. Nursing care records or support records are created, or personal records are created, using various types of record forms. By streamlining the clerical work of record-keeping, the time spent on nursing care records, support records, and personal records can be allocated to the nursing care or support itself.

[0018] In one embodiment, there is provided means for displaying a data-entered record form in which the record content is entered in the data input field of the record item of the record form according to the result of the sorting process received by the above-mentioned receiving means. It is possible to display a data-entered record form as if it had been entered using an input device without necessarily using an input device.

[0019] In one embodiment, the character string based on the above-mentioned utterance includes a character string related to the record content corresponding to the record item and does not include a character string related to the record item for at least one of the plurality of record items included in the above-mentioned record form. That is, the character string (the result of speech recognition) when the record item name is not uttered for at least one of the plurality of record items included in the record form is given to the generative AI. Since the processing request sentence given to the generative AI includes a plurality of record items provided in the record form, the generative AI can accurately grasp the record items for which the record content should be sorted, even if they are not included in the utterance. Therefore, even when the record content for which the record item is not specified during the utterance is included, the generative AI can execute the sorting process of associating the appropriate record content with the appropriate record item. Of course, for some of the plurality of record items included in the record form, during the utterance, not only the record content but also the record item (name) may be included. For example, instead of saying "36 degrees", say "body temperature 36 degrees" ("body temperature" is the record item and "36 degrees" is the record content). The accuracy of the processing of the generative AI for sorting (associating) the record content with the record item can be improved.

[0020] In addition, when a unit (for example, in the above case, "degree" representing temperature) is included, the accuracy of the processing of the generative AI that associates the recording content with the recording items may be improved. For some of the plurality of recording items, a unit may be included during the utterance.

[0021] In one embodiment, the character string based on the above utterance is a sequence of a plurality of recording contents in an order different from the order of the plurality of recording items included in the above recording form. It has been confirmed that the generative AI can distribute the appropriate recording content to the appropriate recording item without uttering the recording content in the same order as the plurality of recording items included in the recording form.

[0022] In one embodiment, the generative AI executes an inference process of inferring the recording content corresponding to the recording items included in the above recording form from the context of the character string based on the above utterance, and executes a process of distributing the inferred recording content to the recording items according to the content. An example is the number of people attending a meeting (for example, 2 people) when the blood relationship of the meeting attendees (for example, the eldest son and his wife) is included during the utterance. Even if the number of people attending the meeting is not uttered although the number of people attending the meeting is included as a recording item, it has been confirmed that the generative AI can accurately infer that the recording content corresponding to the recording item "number of people attending the meeting" is "2 people" because the fact that the meeting attendees are the eldest son and his wife is included in the utterance. By using the generative AI, the inference results by the generative AI can be distributed to the recording items.

[0023] The data content voice input system according to the present invention includes a recording management terminal (client) and a recording management server. The recording management terminal includes a form name of a recording form having a plurality of data items (recording items) and data input fields for each data item, and a microphone that receives an input of speech including data content (recording content) to be input to the data input fields, and voice recognition means for converting the speech including the recording form name and data content received by the microphone into a character string, and transmission means for transmitting the character string converted by the voice recognition means to the recording management server. The recording management server includes reception means for receiving the character string transmitted by the transmission means of the recording management terminal, a plurality of data items included in the recording form specified by the recording form name included in the character string received by the reception means, and data content (which may be the character string itself) included in the character string, and processing request sentence creation means for creating a processing request sentence including a sorting instruction for associating data content with the data items, and transmits the processing request sentence including the data items, data content, and sorting instruction created by the processing request sentence creation means to a generative AI, and generative AI utilization means for receiving, according to the processing request sentence, a sorting result of sorting the data content included in the speech into data items corresponding to the content thereof by the generative AI, and transmission means for transmitting the sorting result received by the generative AI utilization means and form data representing the recording form to the recording management terminal. The recording management terminal further includes creation means for creating data representing a recorded form with data input, in which the data content included in the speech is sorted and input into the data input fields of the data items corresponding to the content thereof according to the sorting result and form data representing the recording form transmitted from the transmission means of the recording management server, and display means for displaying the recorded form with data input according to the data representing the recorded form with data input created by the creation means.

[0024] Data representing a data-entered record form may be created on the record management terminal or on the record management server. When creating on the record management terminal, the distribution processing result and form data representing the record form are sent from the record management server to the record management terminal. When created on the record management server, data representing the data-entered record form is sent from the record management server to the record management terminal.

[0025] Data content is input into the data input fields of the record form according to the speech received by the microphone provided in the record management terminal. It is not always necessary to use an input device (such as a keyboard), and the efficiency of record clerical work can be improved.

[0026] When a record form name that identifies a record form having a plurality of data items is sent from the record management terminal to the record management server, the record management server includes the plurality of data items included in the record form identified by the record form name in the processing request sentence. Since the data item (its candidate) for which the data content during speech should be distributed is accurately transmitted to the generative AI, even if the data item (name) is not included in the speech, the generative AI can accurately perform the association (distribution) of the data item and the data content suitable for it.

[0027] This invention also provides a recording management device. The recording management device according to this invention creates a processing request sentence including a form name of a recording form provided with a plurality of data items (recording items) given according to speech and data input fields for each data item, and a plurality of data items included in the recording form specified by the form name included in the above string, data content included in the above string, and a sorting instruction for associating data content with the above data items according to the string including the form name and the data content to be input in the above data input field. The processing request sentence creating means transmits the processing request sentence including the above data items, data content, and sorting instruction created by the above processing request sentence creating means to the generative AI, and the generative AI utilization means receives the result of the sorting process in which the data content included in the speech is sorted into the data items according to the content, which is executed by the generative AI according to the above processing request sentence. Data representing a recording form with data input can be created using the result of the sorting process. The result of the sorting process may be directly stored in the storage device, or may be displayed on the screen of the recording management terminal before storage, confirmed for errors (corrected correctly if there are errors), and then stored in the storage device.

[0028] This invention also provides a data input method. The data input method according to this invention is such that the processing request sentence creating means creates a processing request sentence including a form name of a recording form provided with a plurality of data items and data input fields for each data item given according to speech, and a plurality of data items included in the recording form specified by the form name included in the above string, data content included in the above string, and a sorting instruction for associating data content with the above data items, and the generative AI utilization means transmits the processing request sentence including the above data items, data content, and sorting instruction created by the above processing request sentence creating means to the generative AI, and receives the result of the sorting process in which the data content included in the speech is sorted into the data items according to the content, which is executed by the generative AI according to the above processing request sentence.

[0029] In one embodiment, the above string is the result of speech recognition processing based on speech.

[0030] Preferably, a record form master database in which the above record form name is associated with and stored together with a plurality of data items (record items) included in the record form specified by the above record form name is connected to the above record management device. As described above, it is possible to accurately convey to the generative AI the data items to which the data content should be allocated, and the accuracy of the allocation process can be improved.

[0031] In one embodiment, the above processing request sentence includes an instruction sentence that instructs the generative AI to operate as a machine that allocates the above data content for each of the above data items.

[0032] In other embodiments, the above processing request sentence includes a constraint sentence that restricts the operation of the above generative AI.

[0033] In other embodiments, the above processing request sentence includes an instruction sentence that instructs the generative AI to support multiple languages.

[0034] The purpose of using the generative AI is to execute a process of allocating (associating) appropriate data content to appropriate data items without giving detailed instructions (commands). By including in the processing request sentence given to the generative AI an instruction sentence that instructs it to operate as a machine that allocates data content for each data item, the accuracy of the allocation process can be improved. Also, by including in the processing request sentence a constraint sentence that restricts the operation of the generative AI, the format of the data representing the data-filled record form, which is the final product using the processing result of the generative AI, can be made to conform to the processing result of the generative AI. Furthermore, by instructing multi-language support, the allocation process can be accurately executed even if the speech (the string given according to the speech) includes a foreign language or dialect.

[0035] Preferably, when the given string includes a specific keyword, the recording management device is provided with control means for functioning as the processing request sentence creation means and the generative AI utilization means. The recording management device can be made to function as a processing device for another task by a specific keyword. For example, if the specific keyword "data storage" is included in the utterance (string), the recording management device is made to operate so as to utilize the above-described generative AI at that time. And if another specific keyword, for example, "user change" is included, the recording management device can be made to operate so as to execute an operation different from the operation of utilizing the generative AI, for example, a user change process (a process of changing the target user of a data input completed recording form).

[0036] This invention also provides an accident report creation support device. The accident report creation support device according to this invention is an accident report creation support device that supports data input to each data input field of an accident report recording form having a plurality of recording items related to an accident and a data input field for each accident recording item, and includes a microphone that receives input of an utterance regarding the accident content to be input to each data input field of the accident report recording form, voice recognition means for converting the utterance regarding the accident content received by the microphone into a string, and a generative AI to which a processing request sentence including the accident recording items included in the accident report recording form, the accident content (which may be the string itself) included in the string, and a distribution instruction for associating the accident content with the accident recording items is given, and receiving means for receiving, in accordance with the processing request sentence, the result of a distribution process in which the accident content included in the utterance is distributed to the accident recording items corresponding to the content. The accident recording items include, for example, the accident occurrence date, the accident occurrence time, the accident occurrence location, the accident type, and the accident occurrence situation. An accident report is created when a user falls down the stairs and gets injured in a care facility.

[0037] The processing request sentence given to the generative AI includes a plurality of accident record items (names) provided in the accident report record form (for example, accident occurrence date, accident occurrence time, accident occurrence location, accident type, and accident occurrence situation). Since the accident record items (names) for classifying the accident content based on the utterance included in the processing request sentence are accurately transmitted to the generative AI, the accuracy of the association (classification) between the accident record items and the accident record content can be improved. Since the accident record content included in the utterance is classified (associated) into the accident record items corresponding to the content, it is possible to input the record content (data content) suitable for the record items into the data input fields for each accident record item provided in the accident report record form. Since it is not necessary to repeat the operation of selecting the accident record item (or its data input field) using the input device (such as a keyboard) and inputting data into the data input field of the selected accident record item, the efficiency of the clerical work for creating the accident report can be improved.

[0038] This invention further provides a meeting report creation support device. The meeting report creation support device according to this invention includes a plurality of record items related to a meeting, and supports data input to each data input field of a meeting report record form having a data input field for each meeting record item. The device includes a microphone that receives an input of an utterance related to the meeting to be input to each data input field of the meeting report record form, voice recognition means for converting the utterance related to the meeting received by the microphone into a character string, and a generative AI to which a processing request sentence including the meeting record items provided in the meeting report record form, the meeting content included in the character string, and a sorting instruction for associating the meeting content with the meeting record items is given. The device is provided with receiving means for receiving the result of a sorting process in which the meeting content included in the utterance is sorted into the meeting record items corresponding to the content according to the processing request sentence. For example, a meeting report is created when a relative of a user of a nursing facility comes to the nursing facility for a meeting. The meeting record items included in the meeting report record form include, for example, the meeting participants, relationship, condition, and number of meeting participants.

[0039] The processing request sentence given to the generative AI includes a plurality of meeting record items (names) (for example, the meeting participants, relationship, situation, and the number of meeting participants) provided in the above meeting report record form. Also, since the meeting record items (names) for classifying the meeting content based on the utterances included in the processing request sentence are accurately transmitted to the generative AI, the accuracy of the association (classification) between the meeting record items and the meeting record content can be improved. Since there is no need to repeatedly select the meeting record items (or their data input fields) using the input device and input data into the data input fields of the selected meeting record items, the efficiency of the clerical work for creating the meeting report can be improved.

Brief Description of the Drawings

[0040]

Figure 1A

Figure 1B

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9A

Figure 9B

Figure 10

Figure 11A

Figure 11B

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Figure 19

Figure 20

Figure 21

Figure 22

Figure 23

Figure 24

Figure 25

Figure 26

Example

[0041] 1. Hardware Configuration FIG. 1A is a block diagram showing the hardware configuration of a voice input system for care content (service content) (hereinafter referred to as the voice input system).

[0042] The voice input system includes a database server 10, a record management server 20, a record management terminal 30, a morphological analysis server 40, and an item allocation server 50. A database (user master database 11, record form master database 12, record table database 13) storing various data necessary for the operation of the voice input system is connected to the database server 10. Other databases can also be connected to the database server 10 as needed.

[0043] The voice input system performs input of various care (support) contents implemented by the staff of a care facility or the like for users of the care facility (typically the elderly or disabled), and records (stores) the input care contents (care record). The care record is used for consideration of a care plan for users of the care facility, utilization for high-quality care services and care management, improvement of the qualifications of the staff of the care facility, information sharing, and the like.

[0044] There are a plurality of care facilities, and generally a plurality of users belong to each of them. Care for each of the plurality of users in the plurality of care facilities is carried out daily, and care records are taken daily. The care record is an administrative task and is often performed by the staff of the care facility who provide care, which is a task separate from the care itself. The voice input system reduces the administrative burden on the staff of the care facility or the like who must perform care records in addition to care.

[0045] In this embodiment, the care content input and recorded using the voice input system includes a number of items (care items) such as the user's vital data (body temperature, blood pressure, percutaneous arterial oxygen saturation, etc.), diet content (amount of food intake), oral care, height, weight, exercise, visitors, recreation content, etc. For each user, the care content is input and recorded (stored). In this embodiment, the care content (care record) includes not only the content related to the daily care of the user but also the user's personal records (such as the above-mentioned body temperature, height, weight, etc.) and reported matters, etc.

[0046] Referring to FIG. 1, generally, a plurality of (or even one) recording management terminals 30 are used for inputting (recording) care content. Although details will be described later, the voice input system of this embodiment can receive the input (including selection) of care content by the speech (voice) of the staff in the care facility, etc., instead of or in addition to the input of care content by text input from the keyboard (including the selection of text by clicking using a mouse). However, even when the care content is input by voice, the care content to be recorded (stored) is text data rather than voice data. In order to support voice input, the recording management terminal 30 is equipped with a voice recognition library 31. The processing of the voice recognition library 31 will be described later.

[0047] The voice input system can be composed of the above-mentioned recording management terminal 30, recording management server 20, and database server 10. In order to enable voice input, a morphological analysis server 40 and an item allocation server 50 are further used. These servers 10, 20, 40, 50 are directly or indirectly connected to the recording management terminal 30 via a network such as the Internet, and the recording management terminal 30 can receive their processing results. The details of the respective processing of the recording management terminal 30 and the servers 10, 20, 40, 50 will be described later.

[0048] FIG. 1B shows the hardware configuration of a voice input system in another aspect.

[0049] The voice input system shown in FIG. 1B integrates the functions of the database server 10 and the recording management server 20 shown in FIG. 1A into the recording management terminal 30. The user master database 11, the recording form master database 12, and the recording table database 13 may be stored in the storage device of the recording management terminal 30, or may be connected to the recording management terminal 30 through a network such as the Internet.

[0050] In the following embodiments, the voice input system shown in FIG. 1A will be described. For the voice input system shown in FIG. 1B, the processing of the database server 10 and the recording management server 20 shown in FIG. 1A is executed in the recording management terminal 30.

[0051] FIG. 2 is a block diagram showing the hardware configuration of the recording management terminal 30. FIG. 3 shows an example of a recording form (health management recording form 32) for inputting (recording) the care content. A plurality of types of recording forms are prepared, and the care content is input according to the care items (service items) defined for each recording form.

[0052] Referring to FIG. 2, the recording management terminal 30 can be used as an operation terminal for inputting care content by voice. The recording management terminal 30 includes a microphone 30G and a display device 30J connected to a data bus 30A. According to the speech of the staff of the care facility collected by the microphone 30G, the care content is input into the recording form displayed on the display device 30J. In the recording form, data input fields for a plurality of different care items (service items) are prepared for each recording form, and care content (numerical values, characters) is input into each of the data input fields.

[0053] The health management record form 32, which is an example of the record form shown in FIG. 3, includes six care items (service items): "body temperature", "blood pressure (upper / lower)", "pulse", "SPO2", "respiration", and "condition observation". For each of the six care items (for blood pressure, both upper and lower), there are data input fields 32a to 32g. Although details will be described later, the voice input system can operate to input appropriate care content based on the speech into the appropriate data input fields for the care items, even if the speech by the staff in the care facility etc. includes words not related to the care items included in the record form, or even if the order of the speech of the care content (numerical values, characters) is different from the order of the care items in the record form, or even if the care item names are not included in the speech.

[0054] Returning to FIG. 2, in addition to the microphone 30G and the display device 30J described above, the data bus 30A is also connected to a CPU 30B that overall controls the operation of the recording management terminal 30, a RAM 30C that temporarily stores data and programs, a storage (such as SSD, HDD) 30D that stores an operating system, various application programs, etc., an input device (such as a keyboard) 30E for inputting characters, numbers, etc., a communication device 30F for transmitting and receiving data through a network, a speaker 30H for outputting sound, and a camera 30I for imaging a subject. A portable computer device capable of network connection can be used as the recording management terminal 30.

[0055] The database server 10, the recording management server 20, the morphological analysis server 40, and the item allocation server 50 are also computers equipped with a CPU, a RAM, a storage, a communication device, etc. However, unlike the recording management terminal 30, since these servers 10, 20, 40, 50 are not computer devices directly operated by the staff in the care facility etc., input devices such as keyboards, display devices, microphones, speakers, and cameras are not necessarily required.

[0056] Section 2. Database Figs. 4 to 6 respectively show the contents of a user master database (hereinafter referred to as the user master DB) 11, a record form master database (hereinafter referred to as the record form master DB) 12, and a record table database (hereinafter referred to as the record table DB) 13 connected to the database server 10.

[0057] Referring to Fig. 4, the user master DB 11 includes data such as a "corporation ID" for identifying the operating corporation of the nursing facility, a "user ID" for identifying a user of the nursing facility, the "name" (in Chinese characters) of the user of the nursing facility specified by the user ID, "gender", the "reading kana" (in katakana) of the name of the user of the nursing facility, and the "reading kana candidates" (in katakana) of the name of the user of the nursing facility. The corporation ID, business operator ID, user ID, name, gender, reading kana, and reading kana candidates, etc. (records) are associated with each other and stored in the user master DB 11.

[0058] The "reading kana candidates" mean candidates for the reading kana of the name assumed from the "name" of the user in Chinese characters. As will be described later, in the voice input system, voice is also used when identifying (selecting) the user for whom the nursing care content should be recorded. By preparing other reading kana candidates in addition to the correct reading kana of the name in the user master DB 11, it is possible to avoid a situation where the user cannot be identified when the name of the user in Chinese characters is not read correctly (not correctly kana-converted).

[0059] Referring to Fig. 5, the record form master DB 12 includes a "corporation ID", a "record form ID" for identifying the record form, the "form name" (in Chinese characters) of the record form, the "reading candidates" (in katakana) of the form name, the "nursing care item name" (service item name) included in the record form, etc.

[0060] "Candidate readings" are candidate readings for identifying a record form. Referring to FIG. 5, for example, the candidate readings for the "Health Management Record Form" with form ID "010011" include "KENKOU" and "VITAL". As will be described later, when "KENKOU" or "VITAL" is included in the speech of the staff of a nursing facility or the like, the voice input system can be operated so that the "Health Management Record Form" with form ID "010011" is identified (selected). The care items included in each of multiple types of record forms differ for each type of record form, and their names are defined as "care item names".

[0061] Referring to FIG. 6, the record table DB13 includes "corporation ID", "operator ID" for identifying a care provider operated by the operating corporation identified by the corporation ID, "user ID", "form ID", "recording date and time", "care content", etc. The care content input into the data input fields for each care item of the above-described record form (for example, the health management record form 32) is stored (saved) in the record table DB13 together with the recording date and time, etc.

[0062] Section 3. Processing of the Voice Input System 3-1. Phase Selection Processing The voice input system performs two major operations: the "user selection phase" and the "record creation phase". Hereinafter, the operation of the voice input system in the "user selection phase" and the operation of the voice input system in the "record creation phase" will be described in detail respectively.

[0063] FIG. 7 is a flowchart showing the flow of the phase selection process for determining whether to operate the voice input system in the "user selection phase" or the "record creation phase". The phase selection process is executed by the record management terminal 30 that constitutes the voice input system.

[0064] When an application program (client program) for operating the computer device as the recording management terminal 30 is started (step 60), a care content input screen (care record screen) is displayed on the display screen of the display device 30J of the recording management terminal 30.

[0065] FIG. 8 shows an example of the care content input screen 33 at startup.

[0066] The care content input screen 33 includes a business operator selection combo 33a, a user selection list 33b, a record form selection combo 33c, a record writing button (save button) 33d, a record form display area 33e, a saved care record list display area 33f, a speech content text display area 33g, a start button 33h, and an end button 33i.

[0067] When a single business operator (care facility) is selected using the business operator selection combo 33a, the names of the users using that business operator are displayed in the user selection list 33b. FIG. 8 shows a state where the names of the users using the business operator "Tokuyo Shizuoka" are displayed in the user selection list 33b because the business operator "Tokuyo Shizuoka" has been selected. When another business operator is selected using the business operator selection combo 33a, the names of the users using that other selected business operator will be displayed in the user selection list 33b. For the display of the names of the business operators displayed in the business operator selection combo 33a and the display of the user names in the user selection list 33b, the above-described user master DB 11 (FIG. 4), etc. are used. That is, when the application program is started on the recording management terminal 30, the recording management terminal 30 and the recording management server 20 are connected by a network (of course, an ID, password, etc. may be required). The recording management server 20 that has received a request from the recording management terminal 30 instructs the database server 10 to read data according to that request. According to the user master DB 11, etc., the business operator names are displayed in the business operator selection combo 33a in a selectable manner, and the user names of the users using the selected business operator are displayed in the user selection list 33b in a selectable manner.

[0068] In the "user selection phase", the user for whom care details should be entered is selected, and in the subsequent "record creation phase", the care details for the user selected in the "user selection phase" are entered. In order to prevent proceeding to the record creation phase without a user being selected, the user may be selected by manual input (clicking using a mouse) rather than voice input only at startup. Of course, the user for whom care details should be entered may be selected by voice input from startup.

[0069] When the start button 33h provided on the care details input screen 33 is clicked, the speech of a caregiver or the like is received by the microphone 30G provided in the record management terminal 30 and input into the record management terminal 30 (step 61).

[0070] The voice data received from the microphone 30G is provided to the voice recognition library 31 (see FIG. 1) provided in the record management terminal 30, where it is converted into text (transcribed) (step 62). The voice recognition library 31 can use a virtual assistant (such as Siri or Google Now) included in the operating system of the record management terminal 30. The voice recognition library 31 converts the speech content into a string of characters mixed with Chinese characters.

[0071] It is assumed that the speech input into the record management terminal 30 includes either the keyword "user change" or "data save" as a keyword. Further, the speech including the keyword "user change" further includes the name (user name), and the speech including the keyword "data save" includes the record form name at the beginning of the speech and also includes the care details.

[0072] The record management terminal 30 for which the start button 33h has been clicked determines whether a specific keyword (the above-mentioned "user change" or "data save") is included in the string of characters converted into text by the voice recognition library 31 (step 63).

[0073] For example, it is assumed that the user says "Mr. Shizuo Okata, user change" and it is converted into text as such by the speech recognition library 31. In this case, the recording management terminal 30 (application program) recognizes the keyword "user change" and proceeds to the user selection phase (user change at step 63, step 64). For example, it is assumed that the user says "health management, 36.5 degrees, data storage" and it is converted into text as such. Then, the recording management terminal 30 recognizes the keyword "data storage" and proceeds to the record creation phase (at step 63, "data storage", step 65). The keywords for proceeding to the user selection phase and the keywords for proceeding to the record creation phase can be arbitrarily set and programmed in the application program in advance. By recognizing the keyword "user change" or "data storage", the recording management terminal 30 can switch between the user selection phase and the record creation phase.

[0074] If neither of the two keywords "user change" and "data storage" is included during the speech (including the case where speech recognition fails), the recording management terminal 30 may prompt the staff of the nursing facility, etc. so that either of the two keywords is included during the speech (recognized by speech recognition) by displaying a message such as "Please say it again." on the display screen of the display device 30J or outputting it as voice from the speaker 30H. The same applies when both "user change" and "data storage" are included in the speech.

[0075] 3-2. User Selection Phase Figure 9A shows the processing flow of the voice input system in the user selection phase. Figure 9B shows an example of the data transmitted and received in the user selection phase.

[0076] When proceeding to the user selection phase, as described above, the name (user name) and the keyword "user change" are spoken and text-converted by the recording management terminal 30. The recording management terminal 30 transmits the text data of the spoken content to the recording management server 20 (step 71, data example 71A).

[0077] The recording management server 20 creates an analysis request including the received user name and transmits it to the morphological analysis server 40 (step 73, data example 73A). The analysis request sentence created in the recording management server 20 and transmitted to the morphological analysis server 40 is, for example, as follows.

[0078] {"messages":[{"role":"system","content":"Please read the names included below in kana.\nMr. Ichiro Shizuoka, user change\n"}],"max_tokens":600,"temperature":0.1,"top_p":0.1,"frequency_penalty":0,"presence_penalty":0}

[0079] This analysis request sentence is an instruction sentence to send back the reading (kana) of the name (user name) included in the character string "Mr. Ichiro Shizuoka, user change" converted from speech to text by the recording management terminal 30.

[0080] The morphological analysis server 40 that has received the analysis request sentence including the user name decomposes the character string (analysis target sentence) "Mr. Ichiro Shizuoka, user change" included in the analysis request sentence into morphemes, and sends back to the recording management server 20 the reading kana of the nouns considered to be names among the nouns obtained by the decomposition (step 75, data example 75A). The morphological analysis server 40 is equipped with a library regarding names and their reading kana, and uses this to select (extract) a name and determine the reading kana of the selected name.

[0081] The record management server 20 that has received the reading kana data (e.g., "SHIZUOKA ICHIROU") of the user name (e.g., "SHIZUOKA ICHIROU") instructs the database server 10 to search for records with the received reading kana data. The database server 10 refers to the user master DB 11 (Figure 4) and searches for records with the received reading kana data. When a record is found by the search, the user ID stored in association with the reading kana data in that record is transmitted to the record management terminal 30 (step 74, data example 74A).

[0082] The user is identified (selected) by the user ID. When the record management terminal 30 receives the user ID, it displays the care content input screen for the user identified by the received user ID (step 72, data example 72A).

[0083] Figure 10 shows the care content input screen 33 of the record management terminal 30 after receiving the user ID. Different from the care content input screen 33 immediately after startup shown in Figure 8, the name of the user identified by the speech (the user with the received user ID, SHIZUOKA ICHIROU) is highlighted in the user selection list 33b, and in the saved care record list display area 33f, the care records performed on the identified user (SHIZUOKA ICHIROU) in the past (even only the most recent) are displayed. The care records stored in the record table DB 13 (Figure 6) are used for the past care records displayed in the care record list display area 33f. It can be seen that the care content input screen 33 is for the user "SHIZUOKA ICHIROU" from the user name highlighted in the user selection list 33b.

[0084] 3-3. Record Creation Phase Figure 11A shows the processing flow of the voice input system in the record creation phase. Figure 11B shows the data example transmitted and received in the record creation phase.

[0085] As described above, when the keyword "data storage" is included in the character string converted by the speech recognition library 31 in the recording management terminal 30 where the application program is being executed, it proceeds to the recording creation phase.

[0086] When proceeding to the recording creation phase, as described above, the recording form name, care content, and keyword "data storage" at the beginning of the utterance have been spoken and converted into text by the recording management terminal 30. The recording management terminal 30 transmits the text data including the recording form name and care content to the recording management server 20 together with the user ID acquired in the above-described user selection phase (step 81, data example 81A).

[0087] The recording management server 20 creates a processing request sentence including the received recording form name and care content (entire utterance text) and transmits it to the item allocation server 50 (step 83, data example 83A). The processing request sentence created by the recording management server 20 in step 83 and transmitted to the item allocation server 50 has the following content, for example.

[0088] {"messages":[{"role":"system","content":"Morphologically analyze the following sentence and return only the first word. Punctuation is not required.\nHealth management, 36 degrees, 98%, good complexion, good physical condition. Data storage\n"}],"max_tokens":800,"temperature":0,"top_p":0.1,"frequency_penalty":0,"presence_penalty":0}

[0089] This processing request sentence is an instruction sentence to send back the first word included in the character string "Health management, 36 degrees, 98%, good complexion, good physical condition. Data storage" converted from the utterance to text by the recording management terminal 30.

[0090] Unlike the analysis request sentence sent to the morphological analysis server 40, the processing request sentence sent to the item sorting server 50 does not include the name of the user of a nursing facility or the like. By not sending both the user name and the nursing details to the same server (not sending the user name to the item sorting server 50), the personal information of the user can be protected.

[0091] Upon receiving a processing request sentence that includes the received record form name and the nursing details, the item sorting server 50 decomposes the character string (sentence to be analyzed) included in the processing request sentence, in this case "Health management, 36 degrees, 98%, good complexion, good physical condition. Data storage", into morphemes, selects the first word in the order of appearance from among the words obtained by the decomposition, and sends the selected word back to the record management server 20 (step 87, data example 87A). If the sentence to be analyzed is "Health management, 36 degrees, 98%, good complexion, good physical condition. Data storage", the record form name "Health management" at the beginning of the utterance will be sent back to the record management server 20.

[0092] Upon receiving the record form name, the record management server 20 instructs the database server 10 to search for records with the received record form name. The database server 10 searches for records with the received record form name by referring to the record form master DB12 (Figure 5). If a record is found, the form ID included in that record is obtained (step 84). If the received record form name is "Health management", the form ID "010011" is obtained (data example 84A).

[0093] In the record form master DB 12 (Fig. 5), the care item names (service item names) included in the record form specified by the form ID are associated with the form ID. The record management server 20 creates a new processing request sentence including the care item names and care details associated with the acquired form ID, and transmits it to the item allocation server 50 (step 85, data example 85A). The processing request sentence created by the record management server 20 in step 85 and transmitted to the item allocation server 50 has the following content, for example.

[0094] {"messages":[{"role":"user","content":"You are a machine that sorts care records for health management by service item. Based on the following three constraints, please transfer the health management care records corresponding to each service item to the right side of the colon and output them.\n# Constraints on health management care records\n· Do not arbitrarily rewrite the content of the care record.\n· If the year (yyyy) in (yyyy / MM / dd) is missing, assume it is 2023.\n· Assume that the word \"today\" is 2023 / 08 / 30.\n· Convert dialects to standard polite language.\n· In the case of languages other than Japanese, convert them to polite Japanese (personal names can remain in other languages).\n# Constraints on service items\n· Do not arbitrarily rewrite the content of the service item.\n# Constraints on the output result\n· If there is a care record that does not apply to any service item, transfer it to the service item that contains the keywords \"condition\", \"state\", \"situation\", or \"content\".\n· Be aware of the unit of the service item.\n· If there is a quotation mark (\"\") on the left side of the colon, extract only what can be read from the care record and copy it to the right side of the colon.\n· If the content inside the quotation mark (\"\") cannot be read from the care record, enter only \"none\" on the right side of the colon.\n· For service items that cannot be read from the care record, enter only \"none\" on the right side of the colon.\n· Do not arbitrarily rewrite the content of the care record and the service item, and do not output the sorting result (do not change the content inside the quotation mark \"\").\n· Return all the service items presented.\n· Always sort the care record by service item.\n# Health management care record\nHealth management, 36 degrees, 98%, good complexion, good physical condition. Data storage\n# Health management service items\nBody temperature (°C):\nBlood pressure upper (mmHg):\nBlood pressure lower (mmHg):\nPulse (times / min):\nSPO2 (%):\nRespiration (times / min):\nCondition observation:\n"}],"max_tokens":800,"temperature":0,"top_p":0,"frequency_penalty":0,"presence_penalty":0}

[0095] The above processing request text includes the following: (A) command sentence, (B) constraint conditions, (C) care content (care record), and (D) care items (service items).

[0096] (A) Command sentence "You are a machine that distributes care records for health management by service item. Based on the following three constraint conditions, please transfer the care records for health management corresponding to each service item to the right side of the colon and output them." By including the operation purpose in the processing request text (command sentence), the generative AI (a specific example will be described later) can be operated as the item distribution server 50. Since the above command sentence includes the record form name "Health Management", the item distribution server 50 should be informed that it should execute the process of distributing the care records (care content) related to health management by service item (care item).

[0097] (B) Constraint conditions (B-1) Constraint conditions for care records of health management · Do not arbitrarily rewrite the content of the care record. · If the yyyy in (yyyy / MM / dd) is missing, assume it is 2023. · Assume that the word "today" is 2023 / 08 / 30. · Convert dialects to standard polite language. · In the case of languages other than Japanese, convert them to polite Japanese (personal names can remain in other languages). "2023" and "2023 / 08 / 30" in the above constraint conditions are based on the machine time of the record management server 20 that creates the processing request text. Specific examples of dialects and languages other than Japanese will be described later. (B-2) Constraint conditions for service items · Do not arbitrarily rewrite the content of the service item. "Service item" means the same as "care item". Whether to use "service item" or "care item" in the processing request sentence may be selected according to, for example, the processing accuracy of the item allocation server 50. (B-3) Constraints on the output result · If there is a care record that does not apply to any service item, transfer it to a service item that contains keywords such as "condition", "state", "situation", and "content". · Be aware of the unit of the service item. · When there is a quotation mark (「」) on the left side of the colon, extract only what can be read from the care record and copy it to the right side of the colon. · When the content inside the quotation mark (「」) cannot be read from the care record, enter only "none" on the right side of the colon. · For service items that cannot be read from the care record, enter only "none" on the right side of the colon. · Do not rewrite the contents of the care record and service items arbitrarily and do not output the allocation result (do not change the content inside the quotation mark 「」). · Return all the presented service items. · Always allocate the care record to the service item. By including the above constraints in the processing request sentence, the generative AI can be specialized to operate as the item allocation server 50, and inappropriate processing can be prevented.

[0098] (C) Care content (care record) Health management, 36 degrees, 98%, good complexion, good physical condition. Data storage This is the full text of the speech. "Data storage" during the speech is not the care content (care record) itself, but a keyword for proceeding to the record creation phase as described above.

[0099] (D) Care items (service items) Body temperature (°C): Blood pressure - upper (mmHg): Blood pressure - lower (mmHg): Pulse (beats / min): SPO2 (%): Respiration (times / min): Status Observation: Here, since the record form name of "Health Management" is included in the processing request sentence, the care items (service items) included in the health management record form 32 (Figure 3), which is one of the multiple record forms, are listed. The listed care items (service items) follow the record form master DB (Figure 12). A plurality of care items (service items) included in the record form specified in the utterance (in the above example, the health management record form) are accurately and completely included in the processing request sentence and transmitted to the item sorting server 50.

[0100] The item sorting server 50 that has received the above-mentioned processing request sentence uses the command sentence, constraint conditions, care content (care record), and care items (service items) included in the processing request sentence to associate the care content with the care items (sorting process), and sends the result to the record management server 20 (step 88, data example 88A). For example, the following data (sorting result data) is sent from the item sorting server 50 to the record management server 20.

[0101] {"id":"chatcmpl-76rGJpqMdVuM87knY7GaP7Sk8e5tv","object":"chat.completion","created":1681869063,"model":"gpt-35-turbo","usage":{"prompt_tokens":856,"completion_tokens":80,"total_tokens":936},"choices":[{"message":{"role":"assistant","content":"Body Temperature (°C): 36.0\nBlood Pressure Upper (mmHg): \nBlood Pressure Lower (mmHg): \nPulse (times / min): \nSPO2 (%): 98\nRespiration (times / min): \nStatus Observation: The complexion is good and the physical condition is good."},"finish_reason":"stop","index":0}]}

[0102] This allocation result includes the following content.

[0103] Body temperature (°C): 36.0 Systolic blood pressure (mmHg): Diastolic blood pressure (mmHg): Pulse rate (beats / min): SPO2 (%): 98 Respiration rate (breaths / min): Status observation: The color is good and the physical condition is good.

[0104] The recording management server 20 that received the allocation result transmits the form data of the recording form (data representing a recording form with no data input) specified by the allocation result and the form ID to the recording management terminal 30 (step 86, data example 86A).

[0105] The recording management terminal 30 that received the allocation result and the form data of the recording form creates a data-inputted recording form by inputting care content (numbers, characters) according to the allocation result into each input field of the recording form, and displays a care content input screen 33 including the data-inputted recording form (step 82, data example 82A).

[0106] FIG. 12 shows the care content input screen 33 of the recording management terminal 30 including the data-inputted recording form. The care content input screen 33 shown in FIG. 10 is different in that the data-inputted recording form is displayed in the recording form display area 33e and the speech content in text is displayed in the speech content text display area 33g.

[0107] If the care content is correctly input into the care item of the recording form displayed in the recording form display area 33e, the save button 33d is clicked, and thereby the care content input into the recording form is saved in the recording table DB13 (FIG. 6). If there is an error in the input care content, the error may be corrected using the input device 30E.

[0108] The voice input ends when the end button 33i is clicked.

[0109] The record form display area 33e of the care content input screen 33 is used to visually confirm whether the care content input by speech is correctly performed (and to correct it if there is an error). Since the speech recognition result of the speech content is also displayed in the speech content text display area 33g, it is also possible to confirm whether the care content input by speech is correctly recognized in the speech content text display area 33g. Therefore, the record form display area 33e does not necessarily have to be displayed. In a mode where the record form display area 33e is not included in the care content input screen 33, by looking at the transcribed speech content displayed in the speech content text display area 33g, if there is an error, the speech including the record form name, care content, and keyword "data storage" can be repeated again.

[0110] FIG. 13 shows the relationship between the above-described speech content (excluding the keyword "data storage" for proceeding to the record creation phase) and the sorting result by the item sorting server 50 (care record using the health management record form 32).

[0111] As described above, the health management record form 32 is specified (selected) from among a plurality of types of record forms by the record form name "health management" included in the speech. By specifying the health management record form 32, the care items (service items) included in the health management record form 32 ("body temperature", "blood pressure up / down", "pulse", "SPO2", "respiration", and "condition observation") are correctly specified based on the record form master DB 12 (FIG. 5). Since this care item (service item) (name) is included in the processing request sentence created in the record management server 20 and transmitted to the item sorting server 50, the item sorting server 50 can accurately convey the care items (service items) to which the care content (care record) should be sorted.

[0112] In the above example of speech (「Health management, 36 degrees, 98%, good complexion, good physical condition.」), the care item (service item) itself is not spoken at all, and thus the care item is not included in the processing request sentence at all. That is, at the stage of speech and processing request sentence, the care item (service item) and the care content (care record) are not associated at all. However, even if the care item and the care content are not associated at all, the item sorting server 50 that has received the processing request sentence can associate the care item and the care content quite accurately (item sorting). That is, in the above example, only 「36 degrees」 is spoken, and the corresponding care item 「body temperature」 is not spoken anywhere before or after 「36 degrees」 (anywhere in the speech), but the item sorting server 50 can correctly associate 「36 degrees」 with the care item 「body temperature」. Similarly, 「98%」 can be correctly associated with the care item 「SPO2」. Furthermore, regarding the speech of 「good complexion, good physical condition.」, it can be correctly associated with the care item 「condition observation」. The meaning of the words that the care item 「body temperature」 has, the normal range of a person's body temperature, and the unit (「degrees」) immediately following the numerical value 「36」, etc., are considered to be used for the inference to sort (associate) 「36 degrees」 to the care item 「body temperature」 among the multiple care items that the health management record form 32 has. Similarly, the meaning of the words that the care item 「SPO2」 (which means oxygen saturation) has, the normal numerical range of a person's SPO2, and the unit (%) immediately following the numerical value 「98」, etc., are considered to be used for the inference to sort 「98%」 to the care item 「SPO2」 among the multiple care items that the multiple health management record forms 32 have.

[0113] The reason why 「good complexion, good physical condition.」 is sorted to the care item 「condition observation」 among the multiple care items that the health management record form 32 has is considered to be that the constraint conditions of the processing request sentence (for example, 「When there is a care record that does not apply to any service item, if it contains keywords such as 「appearance」, 「condition」, 「situation」, 「content」, transfer it to the service item」) have an effect on the correct sorting result.

[0114] The processing request sentence given to the item sorting server 50 includes an instruction sentence "You are a machine that sorts care records for health management by service item" as described above. It is considered that this instruction sentence also contributes to accurately performing the sorting process of the item sorting server 50. Furthermore, as described above, by adding constraint conditions in addition to the above instruction sentence, it is possible to surely reduce sorting errors (mismatches in association).

[0115] The item sorting server 50 can use "ChatGPT", which is a generative AI (artificial intelligence) available via the Internet or the like. That is, by giving a processing request sentence including the above-described instruction sentence, constraint conditions, etc. to ChatGPT, ChatGPT can be operated as the item sorting server 50 to appropriately execute the sorting process. Bard, Phind, and other generative AIs may also be used as the item sorting server 50.

[0116] In the above example, the record management terminal 30 that has received the sorting result and form data from the record management server 20 creates a data-entered record form. However, the record management server 20 may create a data-entered record form, and the data-entered record form created in the record management server 20 may be transmitted to the record management terminal 30.

[0117] FIG. 14 shows another example of the utterance content (shown in parentheses) and the sorting processing result by the item sorting server 50.

[0118] As described above, when the item sorting server 50 receives a processing request sentence including the care item name and care content (care record) from the record management server 20, it performs a process of associating them (sorting process). The example shown in FIG. 14 includes three vital data of systolic blood pressure, diastolic blood pressure, and body temperature as the care content (care record).

[0119] As described above, the health management record form 32 is specified (selected) from among a plurality of types of record forms based on the record form name "health management" included in the utterance. By specifying the health management record form 32, the record form master DB12 is referred to, and the care item names ("body temperature", "blood pressure up / down", "pulse", "SPO2", "respiration", and "condition observation") are correctly specified.

[0120] In the example shown in FIG. 14, different from FIG. 13, the care items (service items) and the care details (care records) are already associated at the utterance stage (the care details corresponding to immediately after the care item name are uttered). It is of course possible to associate the care items and the care details in this way at the utterance stage. However, in the example shown in FIG. 14, the care details are not uttered in the order of the care items included in the record form. For example, in the health management record form 32, the care item "body temperature" is the first care item, but in the utterance example shown in FIG. 14, the care item "blood pressure up" is uttered before the care item "body temperature". Thus, regardless of the order of the care details in the utterance, the item allocation server 50 can allocate the appropriate care details to the appropriate care items. Also, the item allocation server 50 can determine that the process of replacing "36 degrees 5 minutes" with "36.5", and that the pulse, SPO2, respiration, and condition observation are not included in the utterance. Furthermore, even if the care item is not accurately uttered like "down is 75" and there is no unit, it is considered that the item allocation server 50 infers from the context of the utterance that the minimum blood pressure is 75 (mmHg). By including an instruction sentence "You are a machine that allocates health management care records for each service item." in the processing request sentence for the generative AI as described above, and by adding further constraints, the generative AI can be made to operate appropriately as the item allocation server 50, and the errors in allocation (errors in association) can be surely reduced.

[0121] FIG. 15 shows another utterance example and the allocation processing result by the item allocation server 50.

[0122] The word "saturation" is used in this example of speech. The word "saturation" is not included as a care item in the health management record form 32, but is inferred to mean SPO2. Based on the meaning of the word "saturation" and the presence of the unit "%", it is considered that the care content (care record) "98%" is allocated to "SPO2", which is one of the care items included in the health management record form 32.

[0123] Figure 16 shows another example of speech and the result of the allocation process by the item allocation server 50. The example of speech shown in Figure 16 is spoken in English, not Japanese.

[0124] When a staff member engaged in the care service business is a foreigner, speaking in their native language may be able to convey the content more accurately. The above-mentioned processing request sentence includes constraints for multi-language support ("· In the case of a language other than Japanese, convert it to polite Japanese (personal names can remain in other languages)"). This allows the item allocation server 50 to perform the allocation process (including the selection of the record form) after converting the speech to Japanese when the speech is in a foreign language. Even if the speech is in a foreign language, the item allocation server 50 can appropriately select the record form and appropriately associate the care items included in the selected record form with the care content during the speech. In the example of English speech shown in Figure 16, there is no speech about care items, but nevertheless, the care content during the speech can be allocated to appropriate care items.

[0125] Figure 17 shows another example of an example of speech when the speech is in a foreign language (English) and the result of the allocation process by the item allocation server 50. The example of speech shown in Figure 17 is different from Figure 16 in that the care item and the care content are already associated during the speech. Also in this case, the care content can be allocated to appropriate care items.

[0126] FIG. 18 shows another example of speech and the result of the sorting process by the item sorting server 50 according to the example of speech. The processing request sentence created in the record management server 20 and transmitted to the item sorting server 50 is, for example, as follows.

[0127] {"messages":[{"role":"user","content":"You are a machine that sorts out the care records of recreation for each service item. Based on the following three constraints, please transcribe and output the care records of recreation corresponding to each service item to the right of the colon.\n# Constraints on the care records of recreation\n· Do not arbitrarily rewrite the content of the care records.\n· If the yyyy in (yyyy / MM / dd) is missing, assume it is 2023.\n· Assume that the word \"today\" is 2023 / 08 / 30.\n· Convert dialects to standard polite language.\n· In the case of languages other than Japanese, convert them to polite Japanese (personal names can remain in other languages).\n# Constraints on service items\n· Do not arbitrarily rewrite the content of service items.\n# Constraints on the output results\n· If there are care records that do not apply to any service item, transcribe them to the service item that contains the keywords \"condition\", \"state\", \"situation\", \"content\".\n· Be aware of the units of service items.\n· If there are quotation marks (\"\") on the left side of the colon, extract only what can be read from the care records and copy it to the right side of the colon.\n· If the content inside the quotation marks (\"\") cannot be read from the care records, enter only \"none\" on the right side of the colon.\n· For service items that cannot be read from the care records, enter only \"none\" on the right side of the colon.\n· Do not arbitrarily rewrite the content of both the care records and the service items and output the sorting results (do not change the content inside the quotation marks \"\").\n· Return all the service items presented.\n· Always sort the care records into service items.\n# Care records of recreation\nI did recreational gymnastics. I was very enthusiastic, but I got out of breath halfway through.\nPlease do lighter exercises and always take your vital signs from next time."Data storage # Recreation service items Recreation "Implemented or aborted or none": Reason for recreation abortion: Recreation program: Status: Equipment and supplies: Instructions and communication: · Evaluation: "}],"max_tokens}。

[0128] The recreation record form 92 is specified (selected) by "recreation" included in the utterance, and the care items included in the recreation record form 92 ("recreation", "reason for abortion", "program", "status", "equipment and supplies", "instructions and communication", "evaluation") are included in the processing request sentence. It is possible to accurately convey to the item sorting server 50 the care items for which the care content should be sorted.

[0129] In the above-mentioned utterance example, for example, the care item "program" is not uttered, but "gymnastics" during the utterance is correctly sorted into the care item "program". It is considered that the item sorting server 50 can make an inference to sort this into the care item "program" from the meaning of the word "gymnastics". Also, "getting out of breath halfway" during the utterance is sorted into the care item "status". It is considered that the context in the utterance is also used for the sorting process.

[0130] Figure 19 shows still another utterance example and the sorting result by the item sorting server 50 according to the utterance example.

[0131] Since the record form name "moisture" is included in the utterance, the moisture record form 91 is specified (selected) from among a plurality of types of record forms, and the care item names ("beverage name" and "status") included in the moisture record form 91 are specified. It is sorted by the item sorting server 50 that "tea" during the utterance is the beverage name, that the amount of tea consumed is "200 cc" according to "200 cc" immediately after "tea", and that the care content corresponding to "status" is not included in the utterance.

[0132] FIG. 20 shows another example of speech and the result of the sorting process by the item sorting server 50.

[0133] Based on "dinner" included in the speech, the dinner record form 93 is specified (selected), and the names of the care items included in the dinner record form 93 ("swallowing rehabilitation", "amount of food intake", "oral nutrition", "tube nutrition", "medication", "supplementary food", "oral care", "oral care content", "status (meal)", and "skipped meal") are specified. It is determined from the speech "After eating, I brushed my teeth." that oral care has been performed, and the oral care content has been appropriately judged and sorted.

[0134] FIG. 21 shows another example of speech and the result of the sorting process by the item sorting server 50. The processing request sentence created in the record management server 20 and sent to the item sorting server 50 is, for example, as follows.

[0135] {"messages":[{"role":"user","content":"You are a machine that sorts nursing records of physical measurements by service item. Based on the following three constraints, please transcribe and output the nursing records of physical measurements corresponding to each service item to the right of the colon.\n# Constraints on Nursing Records of Physical Measurements\n· Do not arbitrarily rewrite the content of the nursing record.\n· If the yyyy in (yyyy / MM / dd) is missing, assume it is 2023.\n· Assume that the word \"today\" is 2023 / 08 / 30.\n· Convert dialects to standard polite language.\n· In the case of languages other than Japanese, convert to polite Japanese (personal names can remain in other languages).\n# Constraints on Service Items\n· Do not arbitrarily rewrite the content of the service item.\n# Constraints on Output Results\n· If there is a nursing record that does not apply to any service item, transcribe it to the service item that contains the keywords \"condition\", \"state\", \"situation\", \"content\".\n· Be aware of the units of the service items.\n· If there is a quotation mark (\"\") on the left side of the colon, extract only what can be read from the nursing record and copy it to the right side of the colon.\n· If the content inside the quotation mark (\"\") cannot be read from the nursing record, enter only \"none\" on the right side of the colon.\n· For service items that cannot be read from the nursing record, enter only \"none\" on the right side of the colon.\n· Do not arbitrarily rewrite the content of the nursing record and the service item, and do not output the sorting result (do not change the content inside the quotation mark \"\").\n· Return all the service items presented.\n· Always sort the nursing record by service item.\n# Nursing Records of Physical Measurements\nPhysical measurement, 48, 160, The weight has increased slightly. Data storage\n# Service Items of Physical Measurements\nHeight (cm):\nWeight (kg):\nCondition:\n"}],"max_tokens":800,"temperature":0,"top_p":0,"frequency_penalty":0,"presence_penalty":0}

[0136] The body measurement record form 94 is specified (selected) by the "body measurement" included in the utterance, and the care item names ("height", "weight", and "condition") included in the body measurement record form 94 are correctly transmitted to the item allocation server 50.

[0137] Although only "48" and "160" are spoken, "48" is allocated to the care item "weight" and "160" is allocated to the care item "height" respectively. It is considered that the meaning of the words that the care item "weight" has and the normal range of a person's weight are used in the inference for allocating "48" to the care item "weight" among the plurality of care items provided in the body measurement record form 94. Similarly, it is considered that the meaning of the words that the care item "height" has and the normal range of a person's height are used in the inference for allocating "160" to the care item "height" among the plurality of care items provided in the body measurement record form 94. Even if the care content is not spoken in the order of the care items included in the body measurement record form 94, the appropriate care content is allocated to the appropriate care item.

[0138] Figure 22 shows another utterance example and the allocation processing result by the item allocation server 50.

[0139] The care items in the body measurement record form 94 are arranged in the order of height, weight, and condition. Even if the care content is not spoken in that order, it can be seen that the care content of "160 cm" is appropriately allocated to the care item "height" and the care content of "48.3 kg" is appropriately allocated to the care item "weight".

[0140] Figure 23 shows another utterance example and the allocation processing result by the item allocation server 50. The processing request sentence created in the record management server 20 and transmitted to the item allocation server 50 is as follows, for example.

[0141] {"messages":[{"role":"user","content":"You are a machine that distributes the care records of face-to-face meetings according to each service item. Based on the following three constraints, please transcribe the care records of face-to-face meetings corresponding to each service item to the right of the colon and output them.\n# Constraints on the care records of face-to-face meetings\n· Do not arbitrarily rewrite the content of the care records.\n· If the yyyy in (yyyy / MM / dd) is missing, assume it is 2023.\n· Assume that the word \"today\" is 2023 / 08 / 30.\n· Convert dialects to standard polite language.\n· In the case of languages other than Japanese, convert them to polite Japanese (personal names can remain in other languages).\n# Constraints on service items\n· Do not arbitrarily rewrite the content of service items.\n# Constraints on the output result\n· If there is a care record that does not apply to any service item, transcribe it to the service item that contains the keywords \"condition\", \"state\", \"situation\", or \"content\".\n· Be aware of the unit of the service item.\n· If there are quotation marks (\"\") on the left side of the colon, extract only what can be read from the care record and copy it to the right side of the colon.\n· If the content inside the quotation marks (\"\") cannot be read from the care record, enter only \"none\" on the right side of the colon.\n· For service items that cannot be read from the care record, enter only \"none\" on the right side of the colon.\n· Do not arbitrarily rewrite the content of the care record and service item and output the distribution result (do not change the content inside the quotation marks \"\").\n· Return all the provided service items.\n· Always distribute the care record to the service item.\n# Care records of face-to-face meetings\nFace-to-face meeting, the eldest son Taro and his wife, and two grandchildren came for a face-to-face meeting. Well, they were excited talking about the school trip. Data storage\n# Service items for face-to-face meetings\nMeeting attendees:\nRelationship:\nCondition:\nNumber of meeting attendees (person):\n"}],"max_tokens":800,"temperature":0,"top_p":0,"frequency_penalty":0,"presence_penalty":0}

[0142] The meeting record form 95 is identified by the "meeting" included in the utterance, and the item names included in the meeting record form 95 ("meeting participants", "relationship", "status", and "number of meeting participants") are included in the processing request sentence, thereby being correctly transmitted to the item allocation server 50.

[0143] In the above-mentioned utterance example, none of the item names included in the meeting record form 95 are included, but the contents are correctly allocated to the items "meeting participants", "relationship", "status", and "number of meeting participants", respectively.

[0144] Also, the utterance example does not mention at all that the number of meeting participants is 4. The item allocation server 50 correctly determines that the number of meeting participants is 4. This is considered to be the result of correctly inferring that the number of meeting participants is 4 in total, namely "Mr. Taro, the eldest son, his wife, and two grandchildren", according to the part "Mr. Taro, the eldest son, his wife, and two grandchildren" included in the utterance.

[0145] Furthermore, the above-mentioned utterance example also includes an utterance "um~" that has no particular meaning, but the item allocation server 50 also performs a process of deleting this part of the utterance that has no particular meaning (not including it in the item "status").

[0146] Figure 24 shows another utterance example and the allocation result by the item allocation server 50 according to the utterance example.

[0147] Also in the utterance example shown in Figure 24, it is correctly inferred from the part "the eldest son and his wife" in the utterance that the number of meeting participants is 2, and although it is not included in the utterance at all, "2 people" is correctly allocated to the item "number of meeting participants".

[0148] Since it was the "eldest son and his wife" who came to the interview, as the relationship of the interviewees, it would be accurate to say the eldest son and his spouse, but only "eldest son" is allocated in the "relationship" column of the care form 95 shown in Fig. 24. In such a case, in the data input field of the "relationship" column of the care item, using an input device, "spouse of the eldest son", "wife of the eldest son", "daughter-in-law", etc. can be input (added).

[0149] Fig. 25 shows another example of speech and the result of the allocation process by the item allocation server 50.

[0150] The example of speech shown in Fig. 25 contains dialect. As described above, the processing request sentence contains a constraint sentence "Convert dialect to standard polite language", which enables the item allocation server 50 to perform the allocation process after converting the dialect to polite language (standard language). Even if the speech contains dialect, the care items and care contents can be appropriately associated.

[0151] Fig. 26 shows another example of speech and the result of the allocation process by the item allocation server 50 according to the example of speech. The processing request sentence created in the record management server 20 and sent to the item allocation server 50 is, for example, as follows.

[0152] {“messages”:[{“role”:“user”,“content”:“You are a machine that distributes accident report records by service item. Based on the following three constraints, transcribe and output the accident report records corresponding to each service item to the right of the colon.\r\n#Constraints for accident report records\r\n·Do not rewrite the contents of the accident report records\r\n·If the yyyy in (yyyy / MM / dd) is missing, assume that it is 2023\r\n·The word “today” is assumed to be 2023 / 08 / 30\r\n·Convert dialects to standard polite language · In the case of a language other than Japanese, convert it into polite Japanese (person's names can remain in the other language)\r\n#Constraints on service items\r\n·Do not rewrite the contents of the service items without permission\r\n#Constraints on output results\r\n·If there is a record that does not fit into any service item, transcribe it into a service item that contains the keywords "condition", "status", "situation" or "contents"\r\n·Be aware of the unit of service items\r\n·If there are quotation marks (") to the left of the colon, extract only what can be read from the record and copy it to the right of the colon. \r\n·If the contents of the quotation marks ("") cannot be read from the record, enter only "none" to the right of the colon.\r\n·For service items that cannot be read from the record, enter only "none" to the right of the colon.\r\n·Do not arbitrarily rewrite the contents of the record or service items and output the allocation results (do not change the contents of the quotation marks "").\r\n·Return all presented service items.\r\n·Always allocate records to service items.\r\n#Record of accident report\r\nAccident report: Today around 3 p.m., I fell down the stairs and was injured. Umm, it was caused by my careless stepping. Ah, to prevent this from happening again in the future, I need to accompany you as you go down the stairs.Data storage # Service items of accident report Reporter: r Date of occurrence (yyyy / MM / dd): Time of occurrence "AM or PM or none": Time of occurrence (HHmm): Location of occurrence "living room or toilet or dressing room or bathroom or corridor or stairs or conversation room or outside the facility or none": Location of occurrence (outside the facility): Location of occurrence (others): Type of accident "fall or tumble or wrong medicine or burn or accidental ingestion or drowning": Circumstances of accident occurrence: "}],"max_tokens":800,"temperature":0,"top_p":0,"frequency_penalty":0,"presence_penalty":0}。

[0153] The accident report record form 96 is specified by the "accident report" included in the utterance. Although there is no utterance specifying the date in the above example of the utterance, the date "August 30, 2023" is input (assigned) to the item "Date of occurrence" from the "today" included in the utterance. Also, it is determined that it is "PM" from "around 3:00 PM" in the utterance. Furthermore, from "fell from the stairs" in the utterance, the location of occurrence "stairs" and the type of accident "tumble" are selected. Even if there is no utterance for each item of "Date of occurrence", "Time of occurrence", "Location of occurrence", "Type of accident", and "Circumstances of accident occurrence", appropriate contents are assigned to appropriate items.

[0154] The item sorting server 50 using generative AI (ChatGPT) can quite appropriately infer the association (sorting) between the care item (name) and the care content even without uttering the care item name (service item name) and the care content as a set. For example, by clicking on the data input field prepared for each care item on the display screen, the data input field where the care content should be input can be selected, and then it is not necessary to input characters, numbers, etc. using an input device into the selected data input field. Also, it is not necessary to utter the care content in accordance with the order of a plurality of care items included in the record form. The input of the care content can be completed in a short time, and the administrative burden on care staff, etc. will be considerably reduced.

[0155] Note that, although there are data input fields for more appropriate care items, care details may be entered in other data input fields. In such cases, the accuracy of classification can be improved by increasing the above-described constraints.

Explanation of Signs

[0156] 10 Database server 11 User master database 12 Record form master database 13 Record table database 20 Record management server (record management device) 30 Record management terminal (data input support device) 30G Microphone 30J Display device 31 Speech recognition library 32 Health management record form 40 Morphological analysis server 50 Item classification server (generative AI) 91 Moisture record form 92 Recreation record form 93 Dinner record form 94 Physical measurement record form 95 Meeting record form 96 Accident report record form

Claims

1. Analysis request creation means for creating an analysis request including the user name of one user among a plurality of users according to an utterance including the user name of the one user, User identification means for identifying the one user using the analysis result according to the analysis request created by the analysis request means, and Data input screen control means for operating the data input screen as a data input screen for the one user identified by the user identification means. Recording management device.

2. The data input screen includes an area where a record for the one user identified by the user identification means is input, and an area where past records for the one user identified by the user identification means are displayed. The recording management device according to Claim 1.

3. Any one of a plurality of types of record forms is displayed in the area where a record for the one user identified by the user identification means is input. The recording management device according to Claim 2.

4. The user name of the one user identified by the user identification means is highlighted on the data input screen. The recording management device according to Claim 1.

5. The analysis request creation means creates an analysis request including the user name of one user among a plurality of users according to an utterance including the user name of the one user, The user identification means identifies the one user using the analysis result according to the created analysis request, The data input screen control means operates the data input screen as a data input screen for the identified one user. Recording management method.

Images