Information processing method, information processing apparatus, and program

Abstract

Provided is an information processing method including: acquiring a query relating to preparation of a referral letter for a patient, patient information of the patient, and referral-destination attributes that are attributes of a referral-destination doctor; and generating a referral letter corresponding to the referral-destination attributes by inputting the query, the patient information, and the referral-destination attributes to a language model. The referral-destination attributes may be acquired from an attribute database that stores, for each doctor, attributes including a specialty, a career history, and an affiliated clinic.

Description

Information Processing Method, Information Processing Apparatus, and Program

[0001] The present invention relates to an information processing method, an information processing apparatus, and a program.

[0002] Patent Document 1 discloses an information processing system that supports the introduction of patients to a medical institution to be introduced. In this system, a referral is created using information received from a medical institution terminal of the referring medical institution, and a part or all of the general information of the medical institution to be introduced and a part or all of the reservation method information of the medical institution to be introduced are described in a medical guidance ticket.

[0003] Japanese Patent No. 7454305

[0004] Although a system for creating a referral for a medical institution is disclosed as described above, it has been difficult to create a referral that is easy for the institution to be introduced to understand.

[0005] In one aspect of the present disclosure, an information processing method and the like that can create a referral that is easier for the institution to be introduced to understand are provided.

[0006] (1) An information processing method according to one aspect of the present disclosure is an information processing method for obtaining a query regarding the creation of a referral for a patient, patient information of the patient, and an attribute of the doctor at the institution to be introduced (referred-to attribute), and generating a referral according to the referred-to attribute by providing the query, the patient information, and the referred-to attribute to a language model.

[0007] (2) In the information processing method of (1) above, the referred-to attribute may be obtained from an attribute database that stores attributes including the specialty, experience, and affiliated hospital of each doctor, and the query including a creation instruction for the referral according to the referred-to attribute may be provided to the language model.

[0008] (3) In the information processing method of (2) above, a prompt including the creation instruction for the referral, the name of the injury or illness, the referred-to attribute, whether the doctor in charge at the institution to be introduced is a specialist, the availability of using abbreviations, and the referral output format may be provided to the language model.

[0009] (4) In the information processing method described in any of (1) to (3) above, the letter of introduction may be generated by the language model which refers to a database in which the contents of past letters of introduction are stored.

[0010] (5) In the information processing method described in any of (1) to (4) above, the contents of the generated letter of introduction may be output, the contents of the revisions may be accepted, and the letter of introduction and the revisions may be given to the language model to generate the revised letter of introduction.

[0011] (6) In the information processing method described in any of (1) to (5) above, the destination attribute may be output along with the content of the generated letter of introduction.

[0012] (7) In the information processing method described in any of (1) to (6) above, the language model may generate the missing information that is not present in the letter of introduction, along with the letter of introduction, and output the contents of the letter of introduction and the missing information.

[0013] (8) In the information processing method described in any of (1) to (7) above, the multiple items to be described in the referral letter may be displayed, the referral letter items selected by the referring physician from among the multiple items to be described in the referral letter may be received, and the query relating to the creation of the referral letter including the referral letter items may be given to the language model.

[0014] (9) In the information processing method described in any of (1) to (8) above, the referral source attribute, which is an attribute of the referring physician, may be obtained and the referral source attribute may be given to the language model.

[0015] An information processing device in one aspect of the present disclosure comprises a control unit, the control unit acquires a query relating to the creation of a referral letter for a patient, patient information of the patient, and referral destination attributes which are attributes of the referring physician, and provides the query, patient information, and referral destination attributes to a language model, thereby generating a referral letter according to the referral destination attributes.

[0016] One aspect of this disclosure is a program that obtains a query related to the creation of a patient referral letter, patient information of the patient, and referral attributes which are attributes of the referring physician, and provides the query, patient information, and referral attributes to a language model, thereby causing a computer to execute a process to generate a referral letter according to the referral attributes.

[0017] One aspect of this disclosure is that it allows for the creation of referral letters that are easier for the recipient to understand.

[0018] This is an overview diagram of a system relating to one aspect of this disclosure. This is a block diagram showing an example of the configuration of a server device. This is a diagram showing an example of a physician attribute database. This is a block diagram showing an example of the configuration of a terminal device. This is a block diagram showing an example of the configuration of a medical record server device. This is a diagram showing an example of a referral letter database. This is a block diagram showing an example of the configuration of a generation server device. This is a flowchart showing an example of the processing of a server device. This is a diagram showing an example of the referral letter creation start screen. This is a diagram showing an example of the referral letter display screen. This is a diagram showing the first half of the prompt form. This is a diagram showing the second half of the prompt form. This is a diagram showing an example of the referral letter output format referenced in the prompt. This is a flowchart showing an example of the processing of a generation server device. This is a diagram showing an example of the referral letter display screen in a modified example. This is a diagram showing the modification content input screen in the modified example. This is a flowchart showing an example of the processing of a server device in the modified example.

[0019] Hereinafter, an embodiment of a system including an example of an information processing method, information processing device, and program relating to one aspect of this disclosure will be described in detail with reference to the drawings. In the description, similar elements will be denoted by the same reference numerals, and redundant explanations will be omitted as appropriate.

[0020] Figure 1 is an overview diagram of a system relating to one aspect of this disclosure. The system relating to this form of disclosure includes an information processing device 1, an information processing terminal 2, a medical record server device 3, and a generation server device 4, and each device transmits and receives information via a network 5 such as the Internet. The medical record server device 3 has a medical record database 36 (Figure 5) that stores patient medical record information. The generation server device 4 has a language model 45 (Figure 7), such as a Large Language Model (LLM), stored in internal or external storage of the generation server device 4.

[0021] Information processing device 1 is an information processing device that processes, stores, and transmits various types of information. Information processing device 1 may be configured as, for example, a server device, a personal computer, a tablet terminal, or a smartphone. Information processing device 1 may be composed of multiple information processing devices, or it may be configured as one of multiple virtual devices (virtual machines) configured within a single information processing device. In this embodiment of the disclosure, to avoid complexity in the explanation, information processing device 1 will be read as server device 1. However, information processing device 1 is not limited to a server device and may be configured as any of the above-described types of information processing devices.

[0022] The information processing terminal 2 receives information output from the server device 1 via the network 5, accepts user input from the information processing terminal 2, and transmits information related to the input to the server device 1 via the network 5. The information processing terminal 2 can be, for example, a personal computer, smartphone, mobile phone, wearable device, or tablet. In this embodiment of the disclosure, to avoid complexity in the explanation, the information processing terminal 2 will be read and described as the terminal device 2.

[0023] Figure 2 is a block diagram showing an example configuration of the server device 1. The server device 1 can be an information processing device composed mainly of electronic circuits using semiconductor circuit elements. The server device 1 has a control unit 11, a communication unit 12, a reading unit 13, and a storage unit 14. The control unit 11, communication unit 12, reading unit 13, and storage unit 14 are connected to each other via a bus 19 or the like. Note that the server device 1 may have other configurations, such as not having a reading unit 13.

[0024] The control unit 11 can be configured to have one or more processing units such as a CPU (Central Processing Unit), MPU (Micro-Processing Unit), GPU (Graphics Processing Unit), FPGA (Field Programmable Gate Array), DSP (Digital Signal Processor), and a quantum processor. The control unit 11 reads and executes the program (or program product) 16 stored in the storage unit 14.

[0025] The communication unit 12 is a communication module for performing communication-related processing and can send and receive information with terminal devices 2, etc., via the network 5. The reading unit 13 can read portable storage media 15 such as CD (Compact Disc)-ROM, DVD (Digital Versatile Disc)-ROM, and USB memory. The control unit 11 may read the program 16 from the portable storage media 15 via the reading unit 13 and save it to the storage unit 14. Alternatively, the control unit 11 may download the program 16 from another computer via the network 5, etc., and save it to the storage unit 14.

[0026] The storage unit 14 includes volatile storage units such as RAM (Random Access Memory), and non-volatile storage units such as ROM (Read Only Memory), HDD (Hard Disk Drive), and flash memory. As shown in Figure 2, the storage unit 14 stores, for example, a program 16 and a physician attribute database 17. The program 16 is a program executed by the control unit 11. In this embodiment, the program 16 may also be referred to as a program product 16. The physician attribute database 17 is a database that stores a physician's specialty, career history, and affiliated hospital. The physician attribute database 17 is optional.

[0027] Figure 3 shows an example of a physician attribute database 17. As shown in this figure, the physician attribute database 17 has fields for physician identification information, physician name, affiliated medical institution, affiliated department, specialty, alma mater / faculty, and other career information. The physician identification information field is for identifying the physician and stores a unique string of characters for each record. The physician name field stores the physician's name. The affiliated medical institution field stores the name of the medical institution to which the physician belongs. The affiliated department field stores the name of the medical department at the affiliated medical institution. The specialty field stores the medical field in which the physician specializes. The alma mater / faculty field stores the name of the university and faculty the physician attended. The other career information stores hospitals where the physician has previously worked or study abroad experience, etc.

[0028] Figure 4 is a block diagram showing an example configuration of terminal device 2. Here, terminal device 2 can be an information processing device composed mainly of electronic circuits using semiconductor circuit elements. Terminal device 2 may have a control unit 21, a communication unit 22, a reading unit 23, a storage unit 24, a display unit 26, and an input unit 27. The components of the control unit 21, communication unit 22, reading unit 23, storage unit 24, display unit 26, and input unit 27 are connected to each other so as to be able to communicate with one another by a bus 29 or the like.

[0029] The control unit 21 can be configured to have one or more processing units such as a CPU (Central Processing Unit), MPU (Micro-Processing Unit), GPU (Graphics Processing Unit), FPGA (Field Programmable Gate Array), DSP (Digital Signal Processor), and a quantum processor. The control unit 21 reads and executes the program stored in the storage unit 24.

[0030] The communication unit 22 is a communication module for performing communication-related processing and can send and receive information with the server device 1, etc., via the network 5. The reading unit 23 can read portable storage media 25 such as CD (Compact Disc)-ROM, DVD (Digital Versatile Disc)-ROM, and USB memory. The control unit 21 can read programs and / or data from the portable storage media 25 via the reading unit 23 and save them to the storage unit 24. The control unit 21 can also download programs from other computers via the network 5, etc., and save them to the storage unit 24.

[0031] The storage unit 24 includes volatile storage devices such as RAM (Random Access Memory), and non-volatile storage devices such as ROM (Read Only Memory), HDD (Hard Disk Drive), and flash memory. The display unit 26 can be a liquid crystal display or an organic EL (Electro Luminescence) display, etc. The input unit 27 can be an input device such as a keyboard, mouse, touch panel, and camera.

[0032] Figure 5 is a block diagram showing an example configuration of the medical record server device 3. The medical record server device 3 can be an information processing device composed mainly of electronic circuits using semiconductor circuit elements. The medical record server device 3 has a control unit 31, a communication unit 32, a reading unit 33, and a storage unit 34. The control unit 31, communication unit 32, reading unit 33, and storage unit 34 are connected to each other so as to be able to communicate with each other by a bus 39 or the like. The reading unit 33 can read a portable storage medium 35. Except for the contents stored in the storage unit 34, each of these components can be configured the same as the information processing device (server device) 1 in Figure 2, so redundant explanations are omitted. Note that the medical record server device 3 may have other configurations, such as not having a reading unit 33, as in the case of server device 1.

[0033] The storage unit 34 stores the medical record database 36 and the referral letter database 37. The medical record database 36 stores patient medical record information. The medical record information includes information such as the diagnosis entered by the doctor during the examination, and various test results, which are stored along with the date and time. The medical record database 36 usually stores medical record information for patients of one medical institution, but it may be made available to other medical institutions with the patient's consent, etc. Furthermore, the medical record server device 3 may be configured to store medical record information for multiple medical institutions.

[0034] The referral letter database 37 stores the contents of issued referral letters. The referral letter database 37 will be described in detail in the explanation of Figure 6. Alternatively, the medical record server device 3 may not be provided, and the storage unit 14 of the server device 1 may contain the medical record database 36 and the referral letter database 37. In this case, the server device 1 includes the functions of the medical record server device 3.

[0035] Figure 6 shows an example of a referral letter database 37. As shown in this figure, the referral letter database 37 has fields for referral letter identification information, referral date, information of the referring medical institution, information of the referring medical institution, basic patient information, disease name, purpose of referral, symptoms and test results, treatment progress, etc. The referral letter database 37 can thus have fields for common items included in a referral letter. In addition, items other than common items may be stored in a separate field, etc., and the combination of item name and content may be listed in JSON (JavaScript Object Notation) format.

[0036] The "Referral Letter Identification Information" field is for identifying the referral letter and stores a unique string of characters for each record. The "Date of Referral" field stores the date the referral letter was created or sent. The "Referral Medical Institution Information" field stores the name of the referral medical institution. The "Referral Medical Institution Information" field may also have fields other than the medical institution name, such as "Address of the Referral Medical Institution," "Department Name," and "Telephone Number." The "Referring Medical Institution Information" field stores the name of the referring medical institution. The "Referring Medical Institution Information" field may also have fields other than the medical institution name, such as "Address of the Referring Medical Institution," "Department Name," and "Telephone Number."

[0037] The "Patient's Basic Information" section should record the patient's name. This section may also include fields for "Gender," "Date of Birth," "Age," "Telephone Number," and "Occupation." The "Diagnosis" section should record the name of the patient's illness. If there are secondary diagnoses in addition to the primary diagnosis, they may be listed. The "Purpose of Referral" section should record the information requested of the referring physician. The "Symptoms and Test Results" section should record the patient's symptoms and test results that should be useful to the referring institution. The "Treatment Progress" section should record the progression of symptoms and test results at the referring medical institution.

[0038] The fields provided in the referral letter database 37 are not limited to those mentioned above, and any fields can be provided. Furthermore, as described above, items other than the general items may be saved in JSON (JavaScript Object Notation) format, which lists the combination of item name and its content. In addition, as described later, the content of each referral letter may be vectorized and saved in a way that allows searching for similar content. Alternatively, the electronic files of the referral letters may be saved as they are, in which case the referral letter database 37 may save the electronic file name.

[0039] Figure 7 is a block diagram showing an example configuration of the generation server device 4. The generation server device 4 can be an information processing device composed mainly of electronic circuits using semiconductor circuit elements. The generation server device 4 has a control unit 41, a communication unit 42, a reading unit 43, and a storage unit 44. The control unit 41, communication unit 42, reading unit 43, and storage unit 44 are connected to each other so as to be able to communicate with each other by a bus 49 or the like. Except for the contents stored in the storage unit 44, each of these components can be configured the same as the information processing device (server device) 1 in Figure 2, so redundant explanations are omitted. Note that the generation server device 4 may have other configurations, such as not having a reading unit 43, as in the case of server device 1.

[0040] The memory unit 44 can store the language model 45. The language model 45 consists of a program and training data, and the control unit 41 executes the program by referring to the input data and training data input from an external source. The language model 45 may be stored in a storage device or the like outside the generation server device 4, rather than in the memory unit 44.

[0041] The generation server device 4 is an information processing device that handles language models 45, such as large-scale language models (LLM: Large Language Model) like GPT (Generative Pre-trained Transformer, registered trademark), BERT (Bidirectional Encoder Representations from Transformer), and Gemini (registered trademark). The language models 45 may be stored in the storage unit 44 within the generation server device 4, or they may be stored in storage or the like that which is communicated with the generation server device 4.

[0042] The generation server device 4 inputs input data such as images, audio, and text into the language model 45 and generates a response sentence. The language model 45 is a pre-trained machine learning model and can be a large-scale language model such as GPT (Generative Pre-trained Transformer) or BERT (Bidirectional Encoder Representations from Transformer), but other language models may also be used.

[0043] Server device 1 may be connected to generation server device 4 via network 5, or it may be connected to generation server device 4 directly or via a local network without using network 5. Alternatively, server device 1 may have a language model 45 in its storage unit 14 without providing generation server device 4. In this case, server device 1 includes the functions of generation server device 4. Since generation server device 4 is a device that handles language model 45, from the perspective of server device 1, when server device 1 sends prompts, etc., to generation server device 4, it is expressed as server device 1 sending (inputting) to language model 45.

[0044] FIG. 8 is a flowchart showing an example of the processing of the server device 1. As shown in this flowchart, the terminal device 2 transmits a request for the referral creation start screen 51 (FIG. 9) to the server device 1 (step S11). The server device 1 that has received the request for the referral creation start screen 51 transmits referral creation start screen information (step S12), and the terminal device 2 that has received the referral creation start screen information displays the referral creation start screen 51 on the display unit 26 (step S13).

[0045] FIG. 9 is a diagram showing an example of the referral creation start screen 51. As shown in this figure, the referral creation start screen 51 has a patient identification area 511, a doctor identification area 512, an entry item selection area 513, and a referral creation button 514. The patient identification area 511 is an area for identifying the patient who is the subject of the referral, and can be identified by the patient's medical record identification information (hereinafter referred to as "medical record ID (Identification)") or name, etc. When either the medical record ID or the name is input, a candidate for the other of the medical record ID or the name may be displayed. Also, the medical record ID and the name may be input by, for example, having a tag reader read a tag of a medical institution attached to the patient's wrist or the like.

[0046] The doctor identification area 512 is an area for identifying the referring doctor and the referred doctor. In the example of FIG. 9, items for inputting the medical institution name and the doctor name are provided respectively. For example, when the terminal device 2 is used in the referring medical institution, the server device 1 may pre-enter the referring medical institution name and the name of the referring doctor (who is logged in to the server device 1). Also, when moving from the screen displaying the patient's medical record information on the terminal device 2 to the referral creation start screen 51, the content for identifying the patient identification area 511 and the referring doctor using the displayed medical record information may be pre-entered.

[0047] The description item selection area 513 is an area for selecting items to be listed as items in the referral. As shown in FIG. 9, the description item selection area 513 can be used to indicate inclusion in the description items by selecting (checking) the pre-listed description items. The description item selection area 513 displays only selectable items. For essential items such as the patient's name, they will be description items regardless of selection, so they may not be included. Also, for items that are not essential for the referral but are usually provided, they may be displayed in the description item selection area 513 and pre-selected (checked).

[0048] The referral creation button 514 is a button that sends an instruction to create a referral to the server device 1 after the input in the patient identification area 511, the doctor identification area 512, and the description item selection area 513. By the referral creation button 514, the terminal device 2 sends, together with the instruction to create a referral, the patient's medical record ID, name, the names of the referring and referred medical institutions and doctors respectively, and the referral description items selected in the description item selection area 513. Hereinafter, the patient's medical record ID, name, the names of the referring and referred medical institutions and doctors respectively, and the referral description items are referred to as "basic referral information".

[0049] In this way, the server device 1 can display a plurality of description items of the referral on the display unit 26 of the terminal device 2, receive the referral description items, which are the description items selected by the referring doctor among the plurality of description items, and give a query regarding the creation of a referral including the referral description items to the language model 45. Thereby, the user can create a referral limited to only the necessary items.

[0050] Returning to Figure 8, the user of terminal device 2 inputs information according to the contents of the referral letter creation start screen 51, and terminal device 2 receives the entered basic referral letter information (step S14). Having received the basic referral letter information, terminal device 2 sends a referral letter creation instruction, including the basic referral letter information, to server device 1 (step S15). Upon receiving the referral letter creation instruction, server device 1 retrieves the referral destination attribute, which is the attribute information of the referring physician, from the physician attribute database 17, using information that identifies the referring physician as a key (step S16). The attribute information retrieved here may include the physician's specialty, career history, and affiliated hospital information. In addition to the attribute information of the referring physician, the referral source attribute may also be retrieved.

[0051] Next, the server device 1 retrieves patient information from the medical record database 36 of the medical record server device 3, using patient identification information as a key (step S17). The patient information may include any of the following items: the patient's name, gender, date of birth, age, telephone number, occupation, primary diagnosis, secondary diagnosis, medical history, family history, medical examination records, test results, current prescriptions, and next appointment schedule. The patient information may also include contact information such as the address and telephone number of the medical institution that possesses the patient's medical record information.

[0052] Next, server device 1 creates a prompt to generate a referral letter (step S18). The prompt includes a query that instructs the creation of a referral letter, patient information, and referral destination attributes. If server device 1 has obtained the referral source attributes, which are the attributes of the referring physician, the prompt may also include the referral source attributes. Details of the prompt will be described in detail in the explanation of Figures 11 to 13. Next, server device 1 inputs the created prompt into generation server device 4 and uses generation server device 4 to generate a referral letter (step S19).

[0053] As described above, the server device 1 obtains a query related to the creation of a patient's referral letter, the patient's patient information, and the referral destination attribute, which is the attribute of the referring physician. By providing the query, patient information, and referral destination attribute to the language model 45, the server device 1 generates a referral letter that is appropriate to the referral destination attribute. This makes it possible to create a referral letter that is easier for the referring physician to understand. The server device 1 sends the referral letter to the terminal device 2 (step S20), and the terminal device 2 displays a referral letter display screen 52 (Figure 10) including the referral letter on the display unit 26 (step S21).

[0054] Figure 10 shows an example of the referral letter display screen 52. As shown in this figure, the referral letter display screen 52 is a screen on which information such as referral letters received from the server device 1 is displayed on the display unit 26 of the terminal device 2. As shown in Figure 10, the referral letter display screen 52 has a referral letter display area 521, a referral physician information display area 522, and a save button 525.

[0055] The referral letter display area 521 displays the contents of the received referral letter. The referring physician information display area 522 displays the referring physician's attributes. The referring physician attributes may include information on specialty, career history, and affiliated hospital. In this way, the server device 1 can display the referring physician attributes along with the contents of the generated referral letter. This display allows the user to confirm the referring physician's specialty. For example, as shown in Figure 10, if the referring physician's specialty is respiratory medicine and the patient's primary diagnosis is coronary artery disease, the user can recognize from the screen display that the referring physician is not a specialist in the patient's illness.

[0056] Terminal device 2 can accept the operation of the save button 525 and save the referral letter to medical record server device 3 via the network 5. Here, if the referral letter is to be sent electronically to the receiving institution, the referral letter may be sent to the receiving institution. Alternatively, the referral letter may be exchanged by sharing the information of the referral letter saved in medical record server device 3 with the receiving physician. Alternatively, the referral letter saved in medical record server device 3 may be printed using a printer and delivered to the receiving physician by mail or other means.

[0057] Furthermore, the referral letter display screen 52 may not have a referral physician information display area 522, and the server device 1 may not display the referral physician information on the display unit 26 of the terminal device 2.

[0058] Figures 11 and 12 show the first and second halves of the prompt form, respectively. As shown in these figures, the prompt includes items for role, task (query), input, output, prohibitions, important points, and encouragement. However, it is not necessary to include all of these items. The role is an item that describes the role assigned to the language model 45, and in this form it is "You are an excellent medical clerk." The task (query) is an item that instructs the language model 45 to perform the action, and in this form it is an instruction to "create a referral letter for a patient at your hospital who is transferring to another hospital."

[0059] The inputs are the information necessary to generate the output, which is a referral letter. Each input can be extracted from patient information received from the medical record server device 3, referral destination attributes obtained from the physician attribute database 17, and basic referral letter information received from the terminal device 2, all of which are provided along with the prompts. The generation server device 4 uses the language model 45 to assign content to the sub-items listed within the input items based on the patient information, referral destination attributes, and basic referral letter information received along with the prompts, and uses this information to create the output described later.

[0060] Here, among the input items, "Specialty" and "Experience" under "Referral Destination Attributes" can be extracted from the information of the referral destination attribute. The items under "Referral Destination Attributes" are not limited to "Specialty" and "Experience," and may consist of either "Specialty" or "Experience." The input items other than those under "Referral Destination Attributes," such as "Referral Medical Institution Information," "Date of Referral," "Referring Medical Institution Information," "Referring Physician Name," "Patient Basic Information," "Disease Name," "Past Medical History and Family History," "Test Results," "Next Follow-up," and "Current Prescription," can be extracted based on patient information and referral letter basic information.

[0061] Furthermore, the server device 1 may assign and input the content of each sub-item listed within the prompt input items based on patient information, referral destination attributes, and referral letter basic information, and then send the prompt with the input entered to the generating server device 4. In addition, if referring source attributes obtained from the physician attribute database 17 are to be obtained along with the prompt, referring source attributes can also be added to the input items in addition to patient information and referral destination attributes.

[0062] The output (Figure 12) contains the format of the letter of introduction and notes on how to write it. In this form, the letter of introduction format is in a separate file (Figure 13), but the format may also be included in the prompt body. "Prohibitions" and "Important Points" contain notes on the entire process of creating the letter of introduction, and "Encouragement" contains encouragement for the language model 45.

[0063] In the process of creating the referral letter prompt described above (step S18), the server device 1 can compare the disease name included in the patient information with the specialty of the referring physician included in the referral destination attribute and determine whether the disease name is included in the specialty of the referring physician. If the server device 1 determines that the patient's disease name is not included in the specialty of the referring physician, it will write an instruction in the prompt's output field to ensure that abbreviations of that field are not included in the words used in the referral letter.

[0064] Furthermore, if the server device 1 determines that the patient's illness falls within the specialty of the referring physician, it can include a prompt in the prompt's output field indicating that abbreviations for that field may be used in the referral letter. In this way, the server device 1 can retrieve the referral destination attributes from an attribute database that stores attributes including the physician's specialty, career history, and affiliated hospital for each physician, and provide the language model 45 with a query that includes a command to create a referral letter according to the referral destination attributes. This makes it easier to obtain accurate information about the referring physician, and thus provides a referral letter that is easier for the referring physician to understand.

[0065] For example, in the prompt example in Figure 12, it is determined that the patient's illness is not within the referring physician's specialty, and the output states, "Since 'Dr. XX' is not a cardiology specialist, please use the full name instead of an abbreviation," instructing against using abbreviations. If the physician is a specialist, it may state, for example, "Since 'Dr. XX' is a cardiology specialist, it is acceptable to use an abbreviation." For example, the term "coronary artery bypass surgery" could be used in a referral letter from a physician who is not a cardiology specialist, while the abbreviation "CABG" could be used for a referral letter from a physician who is a cardiology specialist.

[0066] Thus, when a patient's illness or injury differs from the specialty of the referring physician, the referral letter can be written using easily understandable terminology, avoiding abbreviations. Conversely, when the patient's illness or injury and the referring physician's specialty are the same, the referral letter can be written concisely using abbreviations.

[0067] Figure 13 shows an example of a referral letter output format referenced by the prompt. As shown in this figure, the referral letter output format has fields for "Information of the referral destination," "Information of the referring source," "Patient information," "Diagnosis," "Purpose of referral," "Medical history and family history," "Progress of illness and test results," "Next follow-up appointment," "Current prescription," and "Remarks." The items to be included in the referral letter can be increased or decreased on the referral letter creation start screen 51 (Figure 9).

[0068] Of the items displayed in Figure 13, "Information about the referral destination," "Information about the referring institution," "Patient information," "Diagnosis," "Medical history and family history," "Test results," "Next follow-up appointment," and "Current prescription" correspond to the information entered in the input, so the information assigned to the input item is transferred. The "Purpose of referral," "Progress of illness," and "Remarks" fields are filled in using the language model 45 based on patient information obtained from the medical record server device 3, prompts, and, if attached, similar referral letters, etc.

[0069] As described above, the prompt includes instructions for creating a referral letter, the name of the illness or injury, whether the referring physician is a specialist, whether abbreviations can be used, and the referral letter output format. Server device 1 can provide the prompt to language model 45. By including content that takes the referring physician into consideration in the prompt, it is possible to create a referral letter that is easier for the referring physician to understand.

[0070] Figure 14 is a flowchart showing an example of the processing performed by the generation server device 4. In step 19 of Figure 8, the server device 1 inputs a prompt to the generation server device 4, thereby inputting a prompt to the language model 45 and creating a letter of introduction. The letter of introduction can be created by referring to similar letters of introduction. In this case, the acquisition of similar letters of introduction may be done using RAG (Retrieval-Augmented Generation) technology, which combines information from an external knowledge base. Furthermore, when the server device 1 creates a prompt, it may add phrases such as "If there are similar letters of introduction, please refer to them when creating the letter of introduction" to the prompt, thereby causing the server device 1 to perform processing that refers to similar letters of introduction.

[0071] For example, as shown in Figure 14, the generation server device 4 determines whether it has received a prompt (step S31). If the generation server device 4 determines that it has not received a prompt (step S31: NO), it repeats the process in step S31. If the generation server device 4 determines that it has received a prompt (step S31: YES), it requests a similar referral letter from the referral letter database 37 of the medical record server device 3 (step S32).

[0072] Similar referral letters can be identified, for example, by matching or similar one or more of the input items of the referral letter to be created, name of illness or injury, and test results. A vectorized referral letter database 37 may be used to determine similarity. Similar referral letters can also be identified by comparing the entirety of similar referral letters with the entirety of their input items.

[0073] The generation server device 4 determines whether it has received a similar referral letter from the medical record server device 3 (step S33). If a similar referral letter has been received (step S33: YES), the generation server device 4 adds a note to the prompt indicating that it will create a referral letter using the similar referral letter as a reference (step S37), inputs the prompt into the language model 45, and retrieves the created referral letter (step S38). If a similar referral letter has not been received (step S33: NO), it inputs the received prompt into the language model 45 and retrieves the created referral letter (step S38).

[0074] Finally, the generation server device 4 transmits the acquired letter of introduction to the server device 1 (step S39). In this way, the server device 1 can generate a letter of introduction using a language model 45 that references a database in which the contents of past letters of introduction are stored. This makes it possible to create a letter of introduction with items and other details that follow those of past letters of introduction. Note that the generation server device 4 may also create a letter of introduction without referring to similar letters of introduction.

[0075] As described above, in this embodiment, since the attributes of the referring physician are used as one of the input pieces of information, it is possible to create a referral letter that is easier for the referring physician to understand.

[0076] Next, a modified version of the above-described embodiment will be explained. In this modified version, the language model 45 detects the missing information in the creation of the letter of introduction. In this case, during the process of creating the letter of introduction prompt (step S18), the server device 1 outputs the missing information separately from the letter of introduction if there is missing information in the creation of the letter of introduction. For example, the server device 1 may write in the output field of the letter of introduction prompt, "If there is missing information, please output the content of the missing information separately from the letter of introduction."

[0077] Furthermore, if there is missing information, the system may not output the letter of introduction, but instead display and output a message indicating that the letter of introduction cannot be created due to missing information, along with the type and content of the missing information, so that the letter of introduction will not be output unless all the information is complete. In this case, for example, server device 1 may write in the output field of the letter of introduction creation prompt, "If there is missing information, do not output the letter of introduction, and output a message indicating that the letter of introduction cannot be created, along with the type and content of the missing information."

[0078] Furthermore, the missing information may not only include missing information about the input items, but also, for example, points out deficiencies (missing items) when compared to similar referral letters. This allows users to become aware of information that should have been included but which they may not have noticed initially.

[0079] Figure 15 shows an example of a referral letter display screen 53 in a modified example. As shown in this figure, the referral letter display screen 53 has a referral letter display area 521, a referral physician information display area 522, a missing information display area 533, a correction content input button 534, and a save button 525. The referral letter display screen 53 differs from the referral letter display screen 52 in Figure 10 in that it has a missing information display area 533 and a correction content input button 534. The referral letter display area 521, the referral physician information display area 522, and the save button 525 are the same as those in the referral letter display screen 52 in Figure 10, so redundant explanations are omitted.

[0080] The missing information display area 533 is an area that displays information when the language model 45 detects a lack of information or other issues. In the example in Figure 15, it displays "There is no information on medications to be prescribed. There is no information on matters to be advised regarding diet and exercise." In this way, the language model 45 can generate missing information along with the referral letter and display the contents of the referral letter and the missing information. This allows the user to create a complete referral letter by making corrections to fill in the missing information. The correction content input button 534 is used when the user wants to correct the contents of the displayed referral letter. Using the correction content input button 534 displays the correction content input screen 54 (Figure 16).

[0081] Figure 16 shows the modification input screen 54 in the modified example. The modification input screen 54 has a text input area 541, a voice input button 542, an additional information (file reference) button 543, a submit modification button 544, and a back button 545. The text input area 541 is an area where text entered using an input unit 27 such as a keyboard is displayed.

[0082] The voice input button 542 is used when voice input is performed via the input unit 27, such as a microphone. The input voice may be converted to text by the terminal device 2 or server device 1 and displayed in the text input area 541. The additional information (file reference) button 543 is a button that displays a selection of an electronic file when additional information is to be added or modified, and the information is provided as an electronic file.

[0083] The "Submit Correction" button 544 sends the entered text, audio, and / or electronic file to the server device 1. The "Back" button 545 returns to the referral letter display screen 53. In this way, the user can input the corrections in text, audio, and / or electronic file format. If an electronic file is added, the system may also provide a text or audio notification of what file was added. The entered corrections are reflected in the referral letter correction prompt, as described later. Although not shown in Figure 16, the content of the referral letter displayed on the referral letter display screen 53 may also be editable directly.

[0084] Figure 17 is a flowchart showing an example of the processing of the server device 1 in the modified example. Figure 17 omits the display of content that overlaps with Figure 8 and shows the process from step S21, "Display the letter of introduction," onwards. As shown in this figure, when the terminal device 2 receives input of a correction instruction from the user on the letter of introduction display screen 53 using the procedure described in Figure 16 (step S51), the terminal device 2 sends the correction instruction to the server device 1 (step S52). When the server device 1 receives the correction instruction, it creates a correction prompt for the letter of introduction (step S54).

[0085] Server device 1 inputs the created correction prompt into the language model 45 of generation server device 4 to create a correction referral letter (step S55). Server device 1 sends the created correction referral letter to terminal device 2 (step S56), and terminal device 2, having received the correction referral letter, displays it on the referral letter display screen 53 (step S57). On the referral letter display screen 53, terminal device 2 determines whether it has received an instruction to correct or save (step S58).

[0086] If it determines that it has received an instruction to correct (step S58: correct), terminal device 2 returns to the process in step S51 and repeats the same process. If it receives an instruction to save (step S58: save), it sends a save instruction to server device 1 (step S59). Upon receiving the save instruction, server device 1 saves the referral letter in the referral letter database 37 of the medical record server device 3 (step S60).

[0087] A letter of introduction revision prompt may, for example, instruct the user to make revisions to a created letter of introduction by adding or deleting items, or it may instruct the user to revise the text or other content in existing items. The prompt may also include a note indicating that additional information provided via electronic files, etc., will be added. Furthermore, a letter of introduction revision prompt may be created by adding revision details to the "Output," "Prohibited Items," and "Important Items" sections (Figure 12) of a prompt related to an already created letter of introduction.

[0088] As described above, this configuration displays the content of the generated letter of introduction, accepts requests for revisions, and provides the letter of introduction and the revisions to the language model 45 to generate a revised letter of introduction. This allows revisions to be made even if the content of the generated letter of introduction is insufficient. In addition, the deficiency display area 533 on the letter of introduction display screen 53 can display deficiencies in the input items and deficiencies when compared with similar letters of introduction, so that users can instantly grasp the information that needs to be added to the letter of introduction and make revisions on the revision input screen 54.

[0089] As described above, in this embodiment as well, since the attributes of the referring physician are used as one of the input pieces of information, it is possible to create a referral letter that is easier for the referring physician to understand.

[0090] In the above-described configuration, the server device 1, terminal device 2, medical record server device 3, and generation server device 4 are described as being composed of different information processing devices, but some or all of them may be composed of a single information processing device.

[0091] In the above configuration, program 16 is executed on server device 1, but part or all of program 16 may be executed on terminal device 2, medical record server device 3, or generation server device 4. Furthermore, program 16 can be referred to as a program product, software, or software product, which may be provided on a recording medium or distributed via a communication network.

[0092] The forms of this disclosure are illustrative in all respects and not restrictive. The scope of the invention is not limited to those shown in the above disclosure but is shown by the claims, and all modifications within the meaning and scope equivalent to the claims are intended.

[0093] Furthermore, the sequences shown in each embodiment described above are not limited, and within a reasonable scope, the order of each processing step may be changed, and multiple processes may be executed in parallel. The processing entity for each process is not limited, and within a reasonable scope, the processing of each device may be executed by other devices.

[0094] The matters described in each embodiment can be combined with each other. Furthermore, the independent claims and dependent claims described in the claims can be combined with each other in any combination, regardless of the form of reference. In addition, although the claims do not use the form of a multi-claim that further references a multi-claim (multi-multi-claim), a form of combination that uses the multi-multi-claim form that references all higher-level claims is also possible.

[0095] 1 Server device (information processing device) 11 Control unit 12 Communication unit 13 Reading unit 14 Storage unit 19 Bus 15 Portable storage medium 16 Program (program product) 17 Doctor attribute database 2 Terminal device (information processing terminal) 21 Control unit 22 Communication unit 23 Reading unit 24 Storage unit 25 Portable storage medium 26 Display unit 27 Input unit 29 Bus 3 Medical record server device 31 Control unit 32 Communication unit 33 Reading unit 34 Storage unit 35 Portable storage medium 36 Medical record database 37 Referral letter database 39 Bus 4 Generation server device 41 Control unit 42 Communication unit 43 Reading unit 44 Storage unit 45 Language model 49 Bus 5 Network 51 Referral letter creation start screen 511 Patient identification area 512 Physician-Specific Area 513 Item Selection Area 514 Referral Letter Creation Button 52 Referral Letter Display Screen 521 Referral Letter Display Area 522 Referral Physician Information Display Area 525 Save Button 53 Referral Letter Display Screen 533 Missing Information Display Area 534 Correction Content Input Button 54 Correction Content Input Screen 541 Text Input Area 542 Voice Input Button 543 Additional Information Button 544 Send Correction Content Button 545 Back Button

Claims

1. An information processing method that obtains a query related to the creation of a referral letter for a patient, patient information of the patient, and referral destination attributes which are attributes of the referring physician, and generates a referral letter according to the referral destination attributes by providing the query, the patient information, and the referral destination attributes to a language model.

2. The information processing method according to claim 1, which retrieves the referral destination attribute from an attribute database that stores attributes for each physician, including the physician's specialty, career history, and affiliated hospital, and provides the language model with the query including an instruction to create the referral letter according to the referral destination attribute.

3. The information processing method according to claim 2, which provides the language model with prompts including instructions for creating the referral letter, the name of the illness or injury, the attributes of the referral destination, whether the attending physician at the referral destination is a specialist, whether abbreviations can be used, and the referral letter output format.

4. The information processing method according to claim 1, which generates the letter of introduction using a language model that references a database in which the contents of past letters of introduction are stored.

5. The information processing method according to claim 1, which outputs the contents of the generated letter of introduction, accepts the contents to be modified, and provides the letter of introduction and the contents to the language model to generate the modified letter of introduction.

6. The information processing method according to claim 1, which outputs the contents of the generated letter of introduction along with the attributes of the person to be introduced.

7. The information processing method according to claim 1, which generates missing information that is not present in the letter of introduction, along with the letter of introduction, using the language model, and outputs the contents of the letter of introduction and the missing information.

8. The information processing method according to claim 1, which displays a plurality of items to be included in the referral letter, receives the referral letter items selected by the referring physician from among the plurality of items to be included in the referral letter, and provides the language model with the query relating to the creation of the referral letter including the referral letter items.

9. The information processing method according to claim 1, which obtains the referral source attribute, which is an attribute of the referring physician, and provides the referral source attribute to a language model.

10. An information processing device comprising a control unit, wherein the control unit obtains a query relating to the creation of a referral letter for a patient, patient information of the patient, and referral destination attributes which are attributes of the referring physician, and provides the query, patient information, and referral destination attributes to a language model to generate a referral letter according to the referral destination attributes.

11. A program that obtains a query related to the creation of a patient's referral letter, the patient's information, and the referral destination attribute, which is the attribute of the referring physician, and provides the query, the patient information, and the referral destination attribute to a language model, thereby causing the computer to execute a process to generate a referral letter according to the referral destination attribute.

Images