Server device, computer program, and medical document creation support method

Abstract

To improve support for the task of creating medical documents based on dialogue between medical professionals and patients.SOLUTION: A server device 12 is capable of communicating with a client device 11, to which information provision is to be performed. A communication interface 121 receives dialogue data DL corresponding to at least a portion of dialogue between a medical professional and a patient. A processor 122 outputs prompt data PR corresponding to a prompt that causes generative artificial intelligence 130 to generate task performance data TE for creating a medical document based on the dialogue data DL. The prompt includes at least one demonstration for creating the medical document from the dialogue.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present disclosure relates to a server device capable of communicating with a client device that provides information. The present disclosure also relates to a computer program executable by a processor installed in the server device. The present disclosure also relates to a medical document creation support method executed by the client device and the server device. [Background technology]

[0002] Patent Document 1 discloses a system for supporting the creation of a medication history book, which is an example of medical documentation, based on a dialogue between a medical professional and a patient. The system is configured to store the contents of medication instructions as a medication history book. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent Publication No. 2021-064161 Summary of the Invention [Problem to be solved by the invention]

[0004] There is a demand for improved support for the task of creating medical documents based on dialogue between medical professionals and patients. [Means for solving the problem]

[0005] One aspect provided by the present disclosure is a server device capable of communicating with a client device to which information is provided, the server device comprising: an interface that accepts interaction data corresponding to at least a portion of an interaction between a healthcare professional and a patient; a processor that outputs prompt data corresponding to a prompt that causes a generative artificial intelligence to generate task performance data for creating at least a portion of a medical document based on the interaction data; It is equipped with The prompts include at least one demonstration for creating the medical document from the interaction.

[0006] One aspect provided by the present disclosure is a computer program executable by a processor in a server device capable of communicating with a client device that provides information, the computer program comprising: When executed, the server device receiving interaction data corresponding to at least a portion of an interaction between a healthcare professional and a patient; outputting prompt data corresponding to a prompt that causes the generative artificial intelligence to generate task performance data for creating at least a portion of a medical document based on the dialogue data; The prompts include at least one demonstration for creating the medical document from the interaction.

[0007] One aspect provided by the present disclosure is a method for supporting medical document creation, which is executed by a client device that executes information provision and a server device that is installed at a location remote from the client device, the method comprising: outputting, from the client device to the server device, interaction data corresponding to at least a portion of an interaction between a healthcare professional and a patient; outputting, from the server device, prompt data corresponding to a prompt that causes the generative artificial intelligence to generate task performance data for creating at least a portion of a medical document based on the dialogue data; The prompts include at least one demonstration for creating the medical document from the interaction.

[0008] According to the configurations of the above-described embodiments, data for creating at least a portion of a medical document is generated based on a dialogue between a medical professional and a patient. This enhances support for the task of creating medical documents based on the dialogue. This allows medical professionals to redirect the effort required for this task to, for example, patient care, thereby enhancing the possibility of addressing the social issue of "shifting from a drug-centered approach to a patient-centered approach." [Brief explanation of the drawings]

[0009] [Figure 1] 1 illustrates an example of the configuration of a document creation support system according to an embodiment. [Figure 2] 2 illustrates the functional configurations of the client device, server device, and AI service providing device of FIG. 1. [Figure 3] 3 illustrates an example of a prompt configuration corresponding to the prompt data of FIG. 2. DETAILED DESCRIPTION OF THE INVENTION

[0010] Examples of embodiments will be described in detail below with reference to the accompanying drawings. In the drawings referred to in the following description, the scale of each illustrated element is changed as necessary to make it recognizable.

[0011] 1 illustrates an example of the configuration of a document creation support system 10 according to an embodiment. The document creation support system 10 is configured for the purpose of supporting the creation of medical documents based on a dialogue regarding medication between a medical professional and a patient.

[0012] As used herein, the term "medical document" refers to a document containing medical information. Examples of such documents include a patient's medication history, a document related to inquiries, a medication information sheet (tracing report), a home medical care report, and a home medical care plan.

[0013] As used herein, the term "dialogue" refers to an exchange of information between a healthcare professional and a patient. Examples of dialogue include a face-to-face conversation at a healthcare facility, a telephone or online consultation, and a text-based question and answer session via email or the like.

[0014] Examples of "healthcare professionals" in this specification include doctors, nurses, pharmacists, care managers, and registered dietitians. The term "patient" used in this specification refers not only to the patient but also to the patient's caregiver. Examples of caregivers include a guardian in the case where the patient is a child, and a caregiver in the case where the patient requires care.

[0015] The term "medical institution" as used herein refers to an institution where medical professionals provide explanations about pharmaceuticals to patients. Examples of medical institutions include hospitals, doctors' offices, clinics, medical offices, pharmacies, drug stores, drug stores, and nursing homes.

[0016] The document creation support system 10 includes a client device 11, a server device 12, and an AI service providing device 13.

[0017] The client device 11 is a device on which information is provided. The client device 11 is used by a medical institution or a patient. Therefore, there may be multiple client devices 11. The client device 11 may be a stationary device installed in a specific location, or may be a portable device that can be carried by a user.

[0018] The server device 12 is installed in a location remote from the client device 11. The client device 11 and the server device 12 are configured to be able to perform two-way data communication via a communication network 20.

[0019] The AI ​​service providing device 13 (hereinafter abbreviated as "service providing device 13") is a device that provides a service using generative artificial intelligence 130. The service providing device 13 is installed in a location remote from the client device 11 and the server device 12. The server device 12 and the service providing device 13 are configured to be able to perform two-way data communication via a communication network 20.

[0020] As used herein, the term "generative artificial intelligence" refers not only to a trained model (computer algorithm) that can generate response data to input data (prompts) that are not included in the training data, but also to software that runs the computer algorithm and hardware that runs the software. The input data can be in a form corresponding to text, audio, images, etc. The response data can also be in a form corresponding to text, audio, images, etc.

[0021] 2, client device 11 includes a communication interface 111. On the other hand, server device 12 includes a communication interface 121. Each of communication interface 111 and communication interface 121 is a hardware interface configured to enable two-way data communication between client device 11 and server device 12.

[0022] Similarly, the service providing device 13 includes a communication interface 131. Each of the communication interface 121 and the communication interface 131 is a hardware interface configured to enable two-way data communication between the server device 12 and the service providing device 13.

[0023] The client device 11 includes a processor 112. The processor 112 is configured to output dialogue data DL from the communication interface 111 to the server device 12. The dialogue data DL corresponds to at least a part of a dialogue regarding medication between a medical professional and a patient.

[0024] The dialogue data DL can be created based on a dialogue acquired through a user interface (not shown). The user interface may be a part of the client device 11 or may be a device independent of the client device 11.

[0025] For example, a conversation between a healthcare professional and a patient input through a microphone as a user interface is converted into text data by a computer application with a transcription function, and at least a portion of the text data can be provided as dialogue data DL.

[0026] As another example, text data corresponding to at least a part of a question and answer exchange conducted via email between a medical professional and a patient can be extracted as the dialogue data DL.

[0027] As another example, text data corresponding to at least a portion of predetermined questions included in a questionnaire form provided by a medical professional and the patient's answers to the questions can be acquired as the dialogue data DL.

[0028] The dialogue data DL output from the client device 11 is received by the communication interface 121 of the server device 12.

[0029] The server device 12 includes a processor 122. The processor 122 is configured to create prompt data PR based on the dialogue data DL received by the communication interface 121. The prompt data PR corresponds to a prompt that causes the generative artificial intelligence 130 of the service providing device 13 to generate task performance data TE. The task performance data TE is data for creating a medical document based on a dialogue corresponding to the dialogue data DL.

[0030] 3 is a schematic diagram illustrating the configuration of a prompt corresponding to the prompt data PR. The prompt includes a first part P1, a second part P2, and a third part P3.

[0031] The first part P1 is a prompt part including information for role definition and instructions. For example, the first part may include a prompt that requests the generative artificial intelligence 130 to behave as a pharmacist and instructs the generative artificial intelligence 130 to create a medication history record in SOAP (Subjective, Objective, Assessment, and Plan) format based on input information. An example of such a prompt is as follows: You are a pharmacist. Below are the statements you and the patient made during medication instructions. Please fill out a medication history in SOAP format based on the content of this conversation.

[0032] The second part P2 is a prompt part including information for providing at least one demonstration to condition the generative artificial intelligence 130 to respond appropriately. Specifically, the demonstration includes inputs of cases similar to the task to be executed by the generative artificial intelligence 130 and ideal outputs for the inputs. An example of such a prompt is as follows: #Pharmacist and patient comments during medication instruction: Pharmacist: Sorry to have kept you waiting. Your medicine is ready. Patient: It took a while again today. Pharmacist: I'm sorry. It takes time to pack everything. Patient: I agree, but I'm glad you've organized it for me, so don't worry about it. Pharmacist: Thank you for saying that. Patient: Thank you. Pharmacist: Today's medication is for 5 weeks, or 35 days. Patient: Yes, that's because of my next appointment. Pharmacist: As usual, all of the medication for after breakfast is packed together in a pack, with 6 types and 7 pills each. Patient: Yes, it should be business as usual. Pharmacist: What did you talk about with the doctor at the hospital today? Patient: There were no particular problems with the blood tests, so I told her that she should just continue as usual. Pharmacist: Do you ever experience shortness of breath, palpitations, or back pain in your daily life? Patient: No, by the way, I've been told that next time I'll have an electrocardiogram for the first time in a while. Pharmacist: I see. It's important to get regular checkups even if you don't have any symptoms. Patient: That's what my doctor told me too. Pharmacist: Is it okay to give the bedtime medication in the sheet form? Patient: Yes, I'm fine. Pharmacist: Have you been taking Myslee sometimes and sometimes not? Patient: Yes, I drink it every day. Pharmacist: I can only give you a 30-day supply of Myslee at a time, so the number of days is 30. Is that okay? Patient: Yes, I have some things that I have been saving up little by little so it's not a problem. Pharmacist: Then please take the one you have. Patient: I understand. Pharmacist: I believe you are taking up to two tablets of Pulsenid and adjusting the dosage accordingly, but how have you been taking it recently? Patient: I usually take one tablet every day, and if I feel like I'm a little constipated, I increase the dosage to two tablets. Pharmacist: Conversely, have you experienced any diarrhea? Patient: It's okay. Pharmacist: Please continue to monitor the situation and make adjustments accordingly. Patient: Yes. Pharmacist: Also, you've been given an additional antibiotic this time. What happened? Patient: I had some symptoms similar to cystitis, so I consulted them and they gave me some medicine. Pharmacist: I see. Are your symptoms things like having to go to the bathroom frequently or feeling like you have residual urine? Patient: Yes. It doesn't hurt when I go to the toilet, but it does feel uncomfortable. Pharmacist: I see. You've been prescribed a 5-day supply, so even if you think your symptoms have subsided, be sure to take the full dose. Patient: Should I stop taking it even if the symptoms subside? Pharmacist: Yes, that's right. To prevent the development of bacteria that are resistant to the medicine, please take the medicine for the full number of days, regardless of whether you have symptoms or not. Patient: That's right. Pharmacist: Also, although this antibiotic is less likely to cause loose stools, some people may still experience loose stools when taking it. If it's not too severe, there's no problem, so I'd like you to continue taking it, but it may be best to reduce the dosage of your usual laxative, Pulsenid, so please keep an eye on it. Patient: I understand. Pharmacist: Is your next appointment on April 30th? Patient: Yes. Pharmacist: I'm sorry to keep you waiting, but if you'd like, we also have a service where you can send your prescription in advance via our app, so please consider it. Patient: Yes, I'll think about it. #Medication history book filled out after the above medication instructions: Subjective information from the patient #1: My usual medication, 5 weeks worth due to my next appointment, which is on 4 / 30. #2: My doctor told me that there were no abnormalities in my blood test results, so I could continue as usual. #3: I don't experience shortness of breath, palpitations, or back pain in my daily life. #4: It's helpful that it's packaged in one package. #5: My next appointment is for my first ECG in a while. #6: I take Myslee every day without any adjustments. I have a little on hand and it's fine until the next dose. #7: Pulsenid is being adjusted (basically 1 tablet, increased to 2 tablets depending on the situation). No diarrhea. #8: I had symptoms of cystitis (frequent need to go to the bathroom, feeling of residual urine, discomfort when urinating) and when I consulted them, they gave me additional medication. Objective information #1: 28->35 days #2: Add Cravit - Analysis and opinions as a pharmacist (Assessment) #1: No change in regular medications. Compliance is maintained through single-dose administration, so continued support is needed. #2: To prevent resistant bacteria, it is necessary to instruct patients to finish taking the antibiotics. #3: It is necessary to adjust the dosage of Prusenid taking into account diarrhea caused by antibiotics. #4: He said he takes Myslee every day, but I'm worried about the remaining medicine. Future plans for the problem #1: Even if your symptoms have subsided, continue taking your medicine without forgetting to do so. #2: Clavit is an antibiotic used to suppress bacterial growth. Taking it on and off can lead to the development of resistant bacteria, which can make the antibiotic less effective. Therefore, even if your symptoms improve, be sure to take it for the full number of days. #3: Antibiotics can make some people more susceptible to diarrhea If it's not too severe, there's no problem, so continue taking it. #4: You may need to take less Pursenide than usual. Please adjust while observing the situation. ·Observation Plan #1: Make sure you've finished your antibiotics #2: Check for severe diarrhea caused by antibiotics #3: Next electrocardiogram test → Check the results #4: Check remaining Myslee medication and medication status

[0033] The third part P3 is a prompt part that includes input information for causing the generative artificial intelligence 130 to actually execute a task. As an example, the input information includes text that has been converted by the prompt function from an actual conversation between a pharmacist and a patient at a medical institution. In other words, the content of the dialogue data DL received from the client device 11 is included in the third part P3.

[0034] 2, the processor 122 is configured to output the created prompt data PR from the communication interface 121 to the service providing device 13. The prompt data PR is received by the communication interface 131 of the service providing device 13.

[0035] The service providing device 13 includes a processor 132. The processor 132 is configured to input the prompt data PR received by the communication interface 131 to the generative artificial intelligence 130.

[0036] As described above, the generative artificial intelligence 130 generates task performance data TE based on the input prompt data PR. The task performance data TE is created so that the medication history book, which is a medical document created based on the tasks included in the prompt data PR, includes the following information: Subjective information from the patient #1: I was told my blood sugar was high again, just like last time. #2: I think I’m using insulin properly, but I end up with excess. #3: I feel better now that it's in a single packet, but I sometimes forget whether I've taken it or not. #4: I have a little bit of my emergency medicine left. #5: I don't feel dizzy or dazed. Objective #1: No prescription changes #2: Insulin injections and emergency medication left over - Analysis and opinions as a pharmacist (Assessment) #1: There is a high possibility that there are issues with compliance due to residual insulin. #2: Single-dose formulation seems to reduce the burden of administration, but administration management is unclear #3: No side effects such as hypoglycemia Future plans for the problem #1: It is important to continue taking insulin properly. Instruct them to mark each injection on a calendar. #2: From next time, we plan to print the date on each package. ·Observation Plan #1: Checking adherence to self-administration #2: Print the date on each subsequent package

[0037] The processor 132 of the service providing device 13 is configured to output the generated task execution data TE from the communication interface 131 to the server device 12. The task execution data TE is received by the communication interface 121 of the server device 12.

[0038] The processor 122 of the server device 12 is configured to transfer the task performance data TE to the client device 11 through the communication interface 121. The transferred task performance data TE is received by the communication interface 111 of the client device 11.

[0039] The processor 112 of the client device 11 is configured to provide a medical document including information corresponding to the task performance data TE through a suitable user interface, such as a display device, a printing device, etc.

[0040] According to the configuration of this embodiment, data for creating at least a portion of a medication history book is generated based on a dialogue between a healthcare professional and a patient regarding medication. This improves the supportability of the task of creating a medication history book based on the dialogue. This allows healthcare professionals to redirect the effort required for this task to, for example, patient care, thereby increasing the likelihood of addressing the social issue of "shifting from medication-centered work to patient-centered work."

[0041] Other examples of prompt portions that may be included in the first portion P1 and the second portion P2 of the prompt illustrated in FIG. 3 are shown below. You are a pharmacist. Below is what you said during medication instruction. The patient responded to your statement, but there is no data on what the patient said. Please try filling out a medication history in SOAP format based on what you said. ---- #Prescription contents Januvia 25mg, 1 tablet once a day after breakfast Bayer aspirin 100mg, 1 tablet once a day after breakfast (Omitted) #Calculation points Medication management guidance fee (no visit within 3 months) Outpatient medication support fee 2 (Omitted) #Pharmacist's comments during medication instruction: Pharmacist: Sorry to keep you waiting, your medicine is ready. Pharmacist: I'm sorry, but it takes time to pack everything up. Pharmacist: Thank you for saying that. Pharmacist: Today's medication is for 5 weeks, or 35 days. Pharmacist: As usual, all of the medication for after breakfast is packed together in a pack, with 6 types and 7 pills each. Pharmacist: What did you talk about with the doctor at the hospital today? Pharmacist: Do you ever experience shortness of breath, palpitations, or back pain in your daily life? Pharmacist: I see. It's important to get regular checkups even if you don't have any symptoms. Pharmacist: Is it okay to give the bedtime medication in the sheet form? Pharmacist: Have you been taking Myslee sometimes and sometimes not? Pharmacist: I can only give you a 30-day supply of Myslee at a time, so the number of days is 30. Is that okay? Pharmacist: Then please take the one you have. Pharmacist: I believe you are taking up to two tablets of Pulsenid and adjusting the dosage accordingly, but how have you been taking it recently? Pharmacist: Conversely, have you experienced any diarrhea? Pharmacist: Please continue to monitor the situation and make adjustments accordingly. Pharmacist: Also, you've been given an additional antibiotic this time. What happened? Pharmacist: I see. Are your symptoms things like having to go to the bathroom frequently or feeling like you have residual urine? Pharmacist: I see. You've been prescribed a 5-day supply, so even if you think your symptoms have subsided, be sure to take the full dose. Pharmacist: That's right. To prevent the development of bacteria that are resistant to the medicine, please take the medicine for the full number of days, regardless of whether you have symptoms or not. Pharmacist: Also, although this antibiotic is less likely to cause loose stools, some people may still experience loose stools when taking it. If it's not too severe, there's no problem, so I'd like you to continue taking it, but it may be best to reduce the dosage of your usual laxative, Pulsenid, so please keep an eye on it. Pharmacist: Is your next appointment on April 30th? Pharmacist: I'm sorry to keep you waiting, but if you'd like, we also have a service where you can send your prescription in advance via our app, so please consider it. #What should be recorded in the medication history book based on the calculation points: Since outpatient medication support fee 2 is being calculated for the first time, patient consent to single-dose packaging must be obtained. In the Assessment section, state that you believe that packaging the medication in a single dose will improve patient adherence. #The thought process for determining the subjective information from the above medication instructions: From the series of comments made by the pharmacist - "I understand you are limiting the dosage to two tablets of Pruzenid and adjusting it accordingly, but how is it being taken recently?", "Have you had any diarrhea or anything like that?", and "In that case, please continue to monitor the situation and adjust the dosage," it can be inferred that the patient's response is "I am currently taking Pruzenid and adjusting the dosage. I am not experiencing any diarrhea symptoms, or they are controllable and not causing any problems," and therefore the following is recorded in the Subject section of the medication history book: "Pruzenid is being adjusted, and there are no problems with diarrhea." (Omitted below) #Medication history book filled out after the above medication instructions: { "Subjective": { "definition": "Patient's subjective symptoms and complaints", "value": [ "My usual medication, due to my next appointment, is for 5 weeks. My next appointment is on 4 / 30." "The blood test results were normal, so the doctor said I could continue as usual." "I don't experience shortness of breath, palpitations, or back pain in my daily life." "It's helpful to have it packaged in one package." "Next time, I'll have an electrocardiogram for the first time in a while." "I take Myslee every day without any adjustments, and I have a little on hand so there's no problem until the next dose." "Purzenid is being adjusted, diarrhea is not a problem", "I had symptoms of cystitis (frequent need to go to the bathroom, feeling of residual urine, discomfort when urinating) and when I consulted them, they gave me additional medication." ] }, (Omitted) "Observation Plan": { "definition": "Items to be passed on next time", "value": [ "Check if you have finished taking your antibiotics", "Check to see if you have severe diarrhea caused by antibiotics," "Next electrocardiogram test → Check results", "Check remaining Myslee medication and medication status" ] } }

[0042] In this example, the demonstration includes only the pharmacist's utterances from the dialogue between the pharmacist and the patient as input, and also includes a demonstration of the process of inferring the patient's utterances from the inputs to create an ideal medication history book.

[0043] In this case, the dialogue data DL output from the client device 11 to the server device 12 can be created by extracting the utterances of the pharmacists who participated in the dialogue. With this configuration, the data amounts of the dialogue data DL and the prompt data PR can be reduced, thereby suppressing an increase in the processing load on each of the processor 122 of the server device 12 and the processor 132 of the service providing device 13.

[0044] In this example, the demonstration includes information on the medication prescribed to the patient and information on the calculated remuneration for the medication. In this case, the processor 112 of the client device 11 creates the dialogue data DL so as to include information on the medication actually prescribed to the patient and information on the calculated remuneration for the medication. This information can be acquired through an appropriate user interface. Note that either the information on the medication prescribed to the patient or the information on the calculated remuneration for the medication may be omitted.

[0045] With this configuration, it is possible to have the generative artificial intelligence 130 of the service providing device 13 execute more complex tasks. In other words, it is possible to improve support for more complex tasks related to creating medication history books.

[0046] In this example, a portion of the demonstration included in the prompt is written in JSON (Java Script Object Notification) format. By writing at least a portion of the prompt data PR in JSON format, it is possible to suppress an increase in the computational load on the processor 132 of the service providing device 13 and improve the stability of the output accuracy of the generative artificial intelligence 130. However, the prompt may also be written in CSV format or XML format.

[0047] At least one of the processor 112 of the client device 11, the processor 122 of the server device 12, and the processor 132 of the service providing device 13, each having the various functions described above, may be realized by a general-purpose microprocessor operating in cooperation with a general-purpose memory. Examples of the general-purpose microprocessor include a CPU, an MPU, and a GPU. Examples of the general-purpose memory include a ROM and a RAM. In this case, the ROM may store a computer program that executes the above-described processes. The ROM is an example of a non-transitory computer-readable medium on which a computer program is stored. The general-purpose microprocessor specifies at least a portion of the program stored in the ROM, expands it on the RAM, and executes the above-described processes in cooperation with the RAM. The computer program may be pre-installed in the general-purpose memory or may be downloaded from an external server device via a communication network and then installed in the general-purpose memory. In this case, the external server device is an example of a non-transitory computer-readable medium on which a computer program is stored.

[0048] At least one of processor 112, processor 122, and processor 132 may be implemented by a dedicated integrated circuit capable of executing the computer program, such as a microcontroller, an ASIC, or an FPGA. In this case, the computer program is pre-installed in a storage element included in the dedicated integrated circuit. The storage element is an example of a computer-readable medium on which a computer program is stored. At least one of processor 112, processor 122, and processor 132 may also be implemented by a combination of a general-purpose microprocessor and a dedicated integrated circuit.

[0049] The configurations described above are merely examples for facilitating understanding of the present disclosure. Each configuration example may be appropriately modified or combined with other configuration examples without departing from the spirit of the present disclosure.

[0050] In the above embodiment, the patient's medication history is written in SOAP format. In this case, the patient's current condition and issues can be communicated in an easy-to-understand manner not only to other pharmacists but also to doctors and nurses. However, the medication history may be written in another format specified by the medical institution. In this case, the content of the demonstration is adjusted so that the medication history can be output in that format.

[0051] In the above embodiment, the prompt data PR created in the server device 12 is output to the service providing device 13, and the task performance data TE created by the generative artificial intelligence 130 is transferred by the server device 12 to the client device 11. However, as illustrated in Fig. 2, the prompt data PR created in the server device 12 may also be output to the client device 11. In this case, the prompt data PR and the task performance data TE are exchanged between the client device 11 and the service providing device 13.

[0052] The configurations listed below also form part of this disclosure. Item 1: A server device capable of communicating with a client device to which information is provided, an interface that accepts interaction data corresponding to at least a portion of an interaction between a healthcare professional and a patient; a processor that outputs prompt data corresponding to a prompt that causes a generative artificial intelligence to generate task performance data for creating at least a portion of a medical document based on the interaction data; It is equipped with the prompts include at least one demonstration for creating the medical document from the dialogue. Server device. Item 2: the dialogue data is created by extracting information provided by medical professionals who participated in the dialogue; the prompts include at least one demonstration of a process for inferring information provided by a patient participating in the dialogue based on the information. Item 1. The server device according to item 1. Item 3: The prompt includes at least one of information regarding a drug prescribed to the patient and information regarding a fee for administering the drug. Item 3. The server device according to item 1 or 2. Item 4: The medical documents include a medical history of the patient who participated in the dialogue. 4. The server device according to any one of items 1 to 3. Item 5: the prompt includes at least one demonstration for filling out the medication history book in SOAP format. Item 4. The server device according to item 4. Item 6: At least a portion of the prompt is written in JSON format. 6. The server device according to any one of items 1 to 5. Item 7: The processor outputs the prompt data to the generative artificial intelligence and transfers the task performance data to the client device. 7. The server device according to any one of items 1 to 6. [Explanation of symbols]

[0053] 11: Client device, 12: Server device, 121: Communication interface, 122: Processor, 130: Generative artificial intelligence, DL: Dialogue data, PR: Prompt data, TE: Task performance data

Claims

1. A server device capable of communicating with a client device to which information is provided, an interface that accepts interaction data corresponding to at least a portion of an interaction between a healthcare professional and a patient; a processor that outputs prompt data corresponding to a prompt that causes a generative artificial intelligence to generate task performance data for creating at least a portion of a medical document based on the interaction data; It is equipped with the prompts include at least one demonstration for creating the medical document from the dialogue. Server device.

2. the dialogue data is created by extracting information provided by medical professionals who participated in the dialogue; the prompts include at least one demonstration of a process for inferring information provided by a patient participating in the dialogue based on the information. The server device according to claim 1 .

3. The prompt includes at least one of information regarding a drug prescribed to the patient and information regarding a dispensing fee for the drug. The server device according to claim 1 .

4. The medical documents include a medical history of the patient who participated in the dialogue. The server device according to claim 1 .

5. the prompt includes at least one demonstration for filling out the medication history book in SOAP format; The server device according to claim 4.

6. At least a portion of the prompt is written in JSON format. The server device according to claim 1 .

7. The processor outputs the prompt data to the generative artificial intelligence and transfers the task performance data to the client device. The server device according to claim 1 .

8. A computer program that is installed in a server device that can communicate with a client device that provides information and that can be executed by a processor, When executed, the server device receiving interaction data corresponding to at least a portion of an interaction between a healthcare professional and a patient; outputting prompt data corresponding to a prompt that causes the generative artificial intelligence to generate task performance data for creating at least a portion of a medical document based on the dialogue data; the prompts include at least one demonstration for creating the medical document from the dialogue. Computer program.

9. A method for supporting medical document creation, which is performed by a client device that provides information and a server device that is installed at a location remote from the client device, comprising: outputting, from the client device to the server device, interaction data corresponding to at least a portion of an interaction between a healthcare professional and a patient; outputting, from the server device, prompt data corresponding to a prompt that causes the generative artificial intelligence to generate task performance data for creating at least a portion of a medical document based on the dialogue data; the prompts include at least one demonstration for creating the medical document from the dialogue. How to assist with medical documentation.

Images