Medical information processing apparatus, medical information processing method, and medical information processing program

Abstract

To provide tag candidates in the annotation work in accordance with thoughts of a patient and the clinical phase for medical records.SOLUTION: A medical information processing apparatus according to an embodiment comprises an acquisition unit, a setting unit, a determination unit, and an output unit. The acquisition unit acquires text information indicating various types of medical records at multiple time points corresponding to medical information of a patient. The setting unit sets instruction sentence information corresponding to an instruction sentence to be input to a large-scale language model on the basis of the text information. The determination unit determines tag candidate information indicating tag candidates to be assigned to the text information on the basis of the instruction sentence information. The output unit outputs the tag candidate information.SELECTED DRAWING: Figure 3

Description

Technical Field

[0001] The embodiments disclosed in this specification and the drawings relate to a medical information processing apparatus, a medical information processing method, and a medical information processing program.

Background Art

[0002] Conventionally, it is known to apply a machine learning model to process text data. In some applications, the language used in the text data to be analyzed is special and domain-specific. For example, there are many technical terms in medicine, and the language used in medicine is often different from the way the same language is used in more general fields. For example, the annotation work of text data is an operation of tagging sentences and paragraphs of text based on pre-defined classification items.

[0003] In the annotation work, for example, the user inputs, and by utilizing a knowledge base or annotated data in which medical terms are accumulated, tag candidates are presented and selected. These annotation works can perform a certain degree of tagging work by utilizing rule-based or machine learning-based NLP (Natural Language Processing) technology. Also, for example, in a large language model, by clearly describing labeling and text processing instruction sentences and requesting them as operations, a certain degree of tagging work is possible.

[0004] However, since these technologies basically assign the same tag to the same sentence or word, it is difficult to perform tagging considering the patient's background. For example, when using it in a large language model, it is necessary to clearly describe labeling and text processing instruction sentences and request them as operations. Also, information that can correctly understand the context included in these text data is required.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0006] One of the problems to be solved by the embodiments disclosed in this specification and the drawings is to provide tag candidates in the annotation work according to the patient's thoughts and the medical treatment phase for the medical records. However, the problems to be solved by the embodiments disclosed in this specification and the drawings are not limited to the above problems. The problems corresponding to the effects of each configuration shown in the embodiments described later can also be regarded as other problems.

Means for Solving the Problems

[0007] The medical information processing apparatus according to the embodiment includes an acquisition unit, a setting unit, a determination unit, and an output unit. The acquisition unit acquires text information indicating a plurality of types of medical records at a plurality of time points corresponding to the medical information of the patient. The setting unit sets instruction sentence information corresponding to an instruction sentence to be input to a large language model based on the text information. The determination unit determines tag candidate information indicating tag candidates to be assigned to the text information based on the instruction sentence information. The output unit outputs the tag candidate information.

Brief Description of the Drawings

[0008]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

DETAILED DESCRIPTION OF THE INVENTION

[0009] Hereinafter, a medical information system according to the present embodiment will be described with reference to the drawings. In the following embodiments, parts denoted by the same reference numerals perform the same operations, and overlapping descriptions will be omitted as appropriate.

[0010] (Embodiment) FIG. 1 is a block diagram showing an example of the configuration of a medical information system 1 according to an embodiment. The medical information system 1 includes a hospital information server 10, a medical information processing apparatus 20, and an operation terminal 30.

[0011] Further, each system and each apparatus are connected to be communicable with each other via a network. Note that the configuration shown in FIG. 1 is an example, and the number of each system and each apparatus may be arbitrarily changed. Also, an apparatus not shown in FIG. 1 may be connected to the network.

[0012] The hospital information server 10 manages various types of information related to patients and the electronic medical records for each patient. The hospital information server 10 is, for example, a HIS (Hospital Information System). The hospital information server 10 is realized by a computer device such as a workstation. The hospital information server 10 stores the electronic medical record in its own storage circuit or the like. The electronic medical record is data having patient information, medical treatment information, etc. of the patient.

[0013] The patient information of the patient includes information contained in the electronic medical record such as patient ID, gender, age, medical history, doctor's findings, medication history, etc., medical images related to the patient obtained in the past, the order (including examination purpose, imaging site, imaging conditions, etc.) used when obtaining the medical image, and examination results related to the patient.

[0014] The medical treatment information of the patient is information stored as structured data or unstructured data in the electronic medical record. Structured data is data such as disease names, symptoms, drug names, etc. that are selected from options, or data with a defined input format such as blood test results and vital data. Unstructured data is data without a defined input format such as text (natural text) written in the free input field or remarks column of the medical treatment record.

[0015] The medical information processing device 20 is a device that provides tag candidates in the annotation work according to the patient's thoughts and medical treatment phases for the medical treatment information of the patient included in the electronic medical record. Details will be described later.

[0016] The operation terminal 30 is a terminal operated by a user such as a doctor or other medical staff. The operation terminal 30 is an example of a medical processing device. The operation terminal 30 is realized by a computer device such as a personal computer or a tablet terminal. The operation terminal 30 is, for example, a terminal for a user to perform an annotation operation on the medical treatment information (medical treatment record) of the patient included in the electronic medical record.

[0017] Next, the configuration of the operation terminal 30 according to this embodiment will be described. FIG. 2 is a block diagram showing an example of the functional configuration of the operation terminal 30 according to this embodiment. The operation terminal 30 includes a communication interface 31, an input interface 32, a display 33, a memory circuit 34, and a processing circuit 35. Note that the configuration included in the operation terminal 30 is not limited to this.

[0018] The communication interface 31 is connected to the processing circuit 35 and controls the transmission and communication of various data performed between each device connected via the network. For example, the communication interface 31 is realized by a network card, a network adapter, a NIC (Network Interface Controller), or the like.

[0019] The input interface 32 is connected to the processing circuit 35, converts an input operation received from a user such as an operator (medical staff) into an electrical signal, and outputs the electrical signal to the processing circuit 35. Specifically, the input interface 32 converts an input operation received from the operator into an electrical signal and outputs the electrical signal to the processing circuit 35.

[0020] The input interface 32 is realized by, for example, a trackball, a switch button, a mouse, a keyboard, a touch pad that performs an input operation by touching an operation surface, a touch screen in which a display screen and a touch pad are integrated, a non-contact input circuit using an optical sensor, and a voice input circuit.

[0021] Note that in this specification, the input interface 32 is not limited to only those including physical operation components such as a mouse and a keyboard. For example, an electrical signal processing circuit that receives an electrical signal corresponding to an input operation from an external input device provided separately from the device and outputs this electrical signal to the control circuit is also included in the examples of the input interface 32.

[0022] The display 33 is connected to the processing circuit 35 and displays various information and various image data output from the processing circuit 35. For example, the display 33 is realized by a liquid crystal display, a CRT (Cathode Ray Tube) display, an organic EL (OELD: Organic Electro Luminescence Display) display, a plasma display, a touch screen, or the like.

[0023] The memory circuit 34 is connected to the processing circuit 35 and stores various data. The memory circuit 34 is an example of a storage unit. The memory circuit 34 stores patient information. Further, the memory circuit 34 stores medical information corresponding to the patient information. Furthermore, the memory circuit 34 stores tag candidate information described later. The memory circuit 34 stores various programs for realizing various functions by being read and executed by the processing circuit 35. For example, the memory circuit 34 is realized by a semiconductor memory element such as a RAM (Random Access Memory) or a flash memory, a hard disk, an optical disk, or the like.

[0024] The processing circuit 35 controls the operation of the entire operation terminal 30. The processing circuit 35 includes, for example, a reception function 351, an input function 352, and a display control function 353. In the embodiment, each processing function performed by the reception function 351, the input function 352, and the display control function 353 is stored in the memory circuit 34 in the form of a program executable by a computer.

[0025] The processing circuit 35 is a processor that reads a program from the memory circuit 34 and executes it to realize the functions corresponding to the respective programs. In other words, the processing circuit 35 in the state of having read each program has each function shown in the processing circuit 35 of FIG. 2.

[0026] In FIG. 2, the receiver function 351, the input function 352, and the display control function 353 are described as being realized by a single processor. However, it is also possible to configure the processing circuit 35 by combining a plurality of independent processors, and each processor realizes the functions by executing a program.

[0027] Also, in FIG. 2, a single storage circuit such as the storage circuit 34 is described as storing programs corresponding to respective processing functions. However, it is also possible to distribute and arrange a plurality of storage circuits, and the processing circuit 35 may be configured to read the corresponding programs from individual storage circuits.

[0028] The term "processor" used in the above description means, for example, a CPU (Central Processing Unit), a GPU (Graphical Processing Unit), or a circuit such as an application specific integrated circuit (ASIC), a programmable logic device (for example, a simple programmable logic device (SPLD), a complex programmable logic device (CPLD), and a field programmable gate array (FPGA)).

[0029] The processor realizes functions by reading and executing the programs stored in the storage circuit 34. Instead of storing the programs in the storage circuit 34, it is also possible to configure the programs to be directly incorporated into the circuit of the processor. In this case, the processor realizes functions by reading and executing the programs incorporated in the circuit.

[0030] The reception function 351 receives tag candidate information from the medical information processing device 20. The reception function 351 is an example of a reception unit. Also, the reception function 751 acquires an electronic medical record from the hospital information server 10. Details of the tag candidate information received from the medical information processing device 20 will be described later.

[0031] The input function 352 inputs information corresponding to the annotation work for the patient information included in the electronic medical record. The input function 352 is an example of an input unit. The input function 352 inputs, for example, tag information corresponding to the annotation work for the patient information included in the electronic medical record.

[0032] The display control function 353 causes the display 33 to display various screens. The display control function 353 is an example of a display control unit. The display control function 353 displays, for example, the tag information received by the reception function 351 or the electronic medical record. The display control function 353 displays, for example, the tag information input by the input function 352. Note that the display control function 353 is not limited to the tag information, the electronic medical record, and the tag information, and may display other information.

[0033] Conventionally, when a user inputs tag information corresponding to an annotation work for patient information included in an electronic medical record, for example, a knowledge base in which medical terms are accumulated, annotated data, a rule base, or NLP technology based on machine learning is utilized to perform a tagging operation. Since these technologies basically assign the same tag to the same sentence or word, it may be difficult to perform tagging considering the patient's background. Therefore, the medical information processing device 20 according to the present embodiment provides tag candidates in the annotation work according to the patient's thoughts and the medical treatment phase.

[0034] Next, the functional configuration of the medical information processing device 20 according to the present embodiment will be described. FIG. 3 is a block diagram showing an example of the functional configuration of the medical information processing device 20 according to the present embodiment. The medical information processing device 20 includes a communication interface 21, a storage circuit 22, and a processing circuit 23. Note that the functional configuration included in the medical information processing device 20 is not limited to this.

[0035] The communication interface 21 is connected to the processing circuit 23 and controls the transmission and communication of various data performed between each device connected via the network. For example, the communication interface 21 is realized by a network card, a network adapter, a NIC (Network Interface Controller), or the like.

[0036] The storage circuit 22 is connected to the processing circuit 23 and stores various data. The storage circuit 22 is an example of a storage unit. The storage circuit 22 stores patient information. Further, the storage circuit 22 stores medical treatment information corresponding to the patient information. Furthermore, the storage circuit 22 stores summary information, instruction sentence information, tag candidates, and tag candidate information, which will be described later. The storage circuit 22 stores various programs for realizing various functions by being read and executed by the processing circuit 23. For example, the storage circuit 22 is realized by a semiconductor memory element such as a RAM (Random Access Memory) or a flash memory, a hard disk, an optical disk, or the like.

[0037] The processing circuit 23 controls the operation of the entire medical information processing device 20. The processing circuit 23 includes, for example, a first acquisition function 231, a second acquisition function 232, a generation function 233, a setting function 234, a determination function 235, and an output function 236. In the embodiment, each processing function performed by the first acquisition function 231, the second acquisition function 232, the generation function 233, the setting function 234, the determination function 235, and the output function 236 is stored in the storage circuit 22 in the form of a program executable by a computer.

[0038] The processing circuit 23 is a processor that reads a program from the storage circuit 22 and executes it to realize the functions corresponding to the respective programs. In other words, the processing circuit 23 in the state of having read each program has each function shown in the processing circuit 23 of FIG. 3.

[0039] In FIG. 3, it has been described that the first acquisition function 231, the second acquisition function 232, the generation function 233, the setting function 234, the determination function 235, and the output function 236 are realized by a single processor. However, it is also possible to configure the processing circuit 23 by combining a plurality of independent processors, and each processor realizes functions by executing a program.

[0040] Also, in FIG. 3, it has been described that a single storage circuit such as the storage circuit 22 stores programs corresponding to respective processing functions. However, it is also possible to distribute and arrange a plurality of storage circuits, and the processing circuit 23 may be configured to read corresponding programs from individual storage circuits.

[0041] The term "processor" used in the above description means, for example, a CPU (Central Processing Unit), a GPU (Graphical Processing Unit), or a circuit such as an application specific integrated circuit (ASIC), a programmable logic device (for example, a simple programmable logic device (SPLD), a complex programmable logic device (CPLD), and a field programmable gate array (FPGA)).

[0042] The processor realizes functions by reading and executing a program stored in the storage circuit 22. Note that instead of storing the program in the storage circuit 22, it may be configured to directly incorporate the program into the circuit of the processor. In this case, the processor realizes functions by reading and executing the program incorporated in the circuit.

[0043] The first acquisition function 231 acquires patient information from the hospital information server 10. The first acquisition function 231 is an example of an acquisition unit. Specifically, the first acquisition function 231 acquires patient information from the hospital information server 10 via a network. The first acquisition function 231 stores the acquired patient information in the storage circuit 22.

[0044] The second acquisition function 232 acquires medical treatment information from the hospital information server 10. The second acquisition function 232 is an example of an acquisition unit. Specifically, the second acquisition function 232 acquires, from the hospital information server 10 via a network, medical treatment information corresponding to the patient information based on the patient information acquired by the first acquisition function 231. For example, the second acquisition function 232 acquires, as medical treatment information regarding the patient, text information indicating a plurality of types of medical treatment records at a plurality of time points corresponding to the medical treatment information of the patient. The second acquisition function 232 stores the acquired medical treatment information in the storage circuit 22.

[0045] The generation function 233 executes natural language processing on the text information and generates summary information corresponding to the text information. The generation function 233 is an example of a generation unit. Specifically, the generation function 233 executes natural language processing on the text information of a plurality of types of medical treatment records at a plurality of time points included in the medical treatment information acquired by the second acquisition function 232, and generates summary information corresponding to the text information. The summary information includes, for example, medical history information corresponding to the patient's medical history, thought information corresponding to the patient's thoughts, and medical treatment phase information corresponding to the patient's medical treatment phase.

[0046] Here, the summary information generated by the generation function 233 will be described with reference to FIG. 4. FIG. 4 is a schematic diagram showing an example of the summary information according to the embodiment. FIG. 4 shows summary information 41. In FIG. 4, the summary information 41 will be described by taking the medical history information corresponding to the patient's medical history as an example.

[0047] The summary information 41 shown in FIG. 4 shows medical history information corresponding to the patient's medical history, for example, medical history regarding current illness, past history, surgical history, and family history. Also, the medical history includes information such as date and time, disease name, treatment method, and medication information regarding medication for each medical history. The medical history information is, in other words, summary information that corresponds medically and biologically.

[0048] Also, the thought information corresponding to the patient's thoughts included in the summary information 41 describes, for example, the patient's external appearance. For example, when the patient's medical history is breast cancer, the patient considers whether to conserve the breast or have a total removal. Also, as a surgical method for removing the tumor, the patient considers whether to have an open surgery with a large wound or an endoscopic surgery with a relatively small wound. The thought information is, for example, information regarding the patient's external changes.

[0049] Next, the medical treatment phase information included in the summary information 41 will be described with reference to FIG. 5. FIG. 5 is a table showing an example of the medical treatment phase information according to the embodiment. The medical treatment phase information 51 includes, for example, the medical treatment content and the medical treatment phase corresponding to the medical treatment content. The medical treatment content includes, for example, diseases, recurrence, examination, notification·IC (Informed Consent), hospitalization, surgical history, radiation, drug therapy, hormone therapy, genetic testing, etc.

[0050] Returning to FIG. 3, the setting function 234 sets command information based on the text information. The setting function 234 is an example of a setting unit. Specifically, the setting function 234 sets command information corresponding to the command sentence input to the large language model based on the text information of a plurality of types of medical treatment records at a plurality of time points included in the medical treatment information acquired by the second acquisition function 232. For example, the setting function 234 sets command information corresponding to the command sentence input to the large language model based on the summary information 41 generated by the generation function 233.

[0051] Here, the instruction sentence information set by the setting function 234 will be described with reference to FIG. 6. FIG. 6 is a schematic diagram showing an example of the instruction sentence information according to the embodiment. FIG. 6 shows instruction sentence information 61. In FIG. 6, the instruction sentence information 61 describes an instruction sentence based on the thought information corresponding to the patient's thought.

[0052] The instruction sentence information 61 shown in FIG. 6 is instruction sentence information including the thought information corresponding to the patient's thought. The instruction sentence information 61 includes tags corresponding to the thought information corresponding to the patient's thought as an instruction sentence input to a known large language model. In the case of the instruction sentence information 61 shown in FIG. 6, the setting function 234 sets, for example, an instruction sentence such as "Please enclose the words related to hobbies (thought information) <tag1>< / tag1> with [ ]. That is, the instruction sentence information 61 includes tags corresponding to the summary information 41. For example, the instruction sentence information 61 is information including tags corresponding to the medical history information, the thought information, and the medical treatment phase information 51.

[0053] Returning to FIG. 3, the determination function 235 determines tag candidate information indicating tag candidates for attaching to the text information based on the instruction sentence information 61. The determination function 235 is an example of a determination unit. Specifically, the determination function 235 determines tag candidate information indicating tag candidates for attaching to the text information of a plurality of types of medical treatment records at a plurality of time points corresponding to the patient's medical treatment information based on the instruction sentence information 61 set by the setting function 234.

[0054] Here, the tag candidates will be described with reference to FIG. 7. FIG. 7 is a schematic diagram showing an example of the tag candidates according to the embodiment. FIG. 7 shows tag candidates 71 in the annotation work. The tag candidates 71 are tag candidates corresponding to the medical history information corresponding to the patient's medical history, the thought information corresponding to the patient's thought, and the medical treatment phase information 51 corresponding to the patient's medical treatment phase, respectively. Note that the tag candidates are not limited to this, and also include tags used in the annotation work.

[0055] Next, the tag candidate information will be described with reference to FIG. 8. FIG. 8 is a schematic diagram showing an example of the tag candidate information according to the embodiment. FIG. 8 shows tag candidate information 81 determined by the determination function 235. The tag candidate information 81 includes text information of a plurality of types of medical records, tag candidates, and check boxes 82 for the user to select. For example, as shown in FIG. 8, the determination function 235 determines tag candidates 831 and 832 for the text information 821 of the medical record. The determination function 235 determines tag candidates 851 and 852 for the text information 841 of the medical record.

[0056] Also, the tag candidate 831 is a tag candidate associated with medical history information corresponding to the patient's medical history. Further, the tag candidates 832 and 852 are tag candidates associated with thought information corresponding to the patient's thoughts. The tag candidate 851 is a tag candidate associated with the medical phase information 51 corresponding to the patient's medical phase. The determination function 235 determines tag candidate information 81 indicating tag candidates to be assigned to the text information of a plurality of types of medical records at a plurality of time points.

[0057] Returning to FIG. 3, the output function 236 outputs the tag candidate information 81. The output function 236 is an example of an output unit. Specifically, the output function 236 outputs the tag candidate information 81 determined by the determination function 235. Thereby, the medical information processing apparatus 20 according to the present embodiment provides tag candidates in the annotation work according to the patient's thoughts and medical phases.

[0058] Next, the processing executed by the medical information processing apparatus 20 according to the embodiment will be described. FIG. 9 is a flowchart showing an example of the processing content of the medical information processing apparatus 20 according to the embodiment. In the following, for the sake of specific explanation, it will be described assuming that the hospital information server 10 stores the electronic medical records necessary for the annotation work.

[0059] First, the first acquisition function 231 acquires patient information from the hospital information server 10 (step S91). Subsequently, the second acquisition function 232 acquires medical information corresponding to the patient information from the hospital information server 10 based on the patient information acquired by the first acquisition function 231 (step S92). Subsequently, the generation function 233 performs natural language processing on the text information of a plurality of types of medical records at a plurality of time points included in the medical information acquired by the second acquisition function 232, and generates summary information 41 corresponding to the text information (step S93).

[0060] Subsequently, the setting function 234 sets command information 61 corresponding to a command sentence to be input to the large language model based on the summary information 41 generated by the generation function 233 (step S94). Subsequently, the determination function 235 determines tag candidate information 81 indicating tag candidates to be assigned to the text information of a plurality of types of medical records at a plurality of time points based on the command information 61 set by the setting function 234 (step S95). Subsequently, the output function 236 outputs the tag candidate information 81 determined by the determination function 235 (step S96). When this process ends, the process of the processing circuit 23 ends.

[0061] As described above, the medical information processing apparatus 20 according to the embodiment acquires text information indicating a plurality of types of medical records at a plurality of time points corresponding to the medical information of the patient, performs natural language processing on the text information, generates summary information 41 corresponding to the text information, sets command information 61 corresponding to a command sentence to be input to the large language model based on the summary information 41, determines and outputs tag candidate information 81 indicating tag candidates to be assigned to the text information based on the command information 61.

[0062] According to at least one embodiment described above, the medical information processing apparatus 20 outputs tag candidate information 81 indicating tag candidates for attaching to text information indicating a plurality of types of medical records at a plurality of time points corresponding to a patient's medical treatment information, thereby providing tag candidates in the annotation work according to the patient's thoughts and medical treatment phases for the medical records. Further, since the user performs the annotation work while checking the output tag candidate information 81, this leads to a reduction in the work load related to the annotation work performed by the user. Furthermore, since the medical information processing apparatus 20 outputs the tag candidate information 81 according to the patient, the user can perform an annotation work unique to the individual patient.

[0063] In addition, since the user can perform an annotation work unique to the individual patient, it can be used to create a Patient Journey that shows the flow of the entire series of actions related to the patient's recognition of a disease or symptom, visiting a medical institution, and subsequent medication, treatment, etc. using the medical treatment information for which the annotation work has been performed.

[0064] Note that the above-described embodiment can be appropriately modified and implemented by changing a part of the configuration or function of each device. Therefore, some modification examples according to the above-described embodiment will be described below as other embodiments. In the following, the points different from the above-described embodiment will be mainly described, and the same reference numerals will be given to the points common to the already described content, and the detailed description will be omitted. Further, the other embodiments described below may be implemented individually or in appropriate combination.

[0065] (First Modification Example) In the above-described embodiment, the processing circuit of the medical information processing apparatus may acquire feedback from the user regarding the tag candidate information 81 output by the output function 236. FIG. 10 is a block diagram showing an example of the functional configuration of the operation terminal 300 according to the first modification. The processing circuit 350 of the operation terminal 300 according to the first modification includes, for example, a reception function 351, an input function 352, a display control function 353, and a transmission function 354. In the first modification, each processing function performed by the reception function 351, the input function 352, the display control function 353, and the transmission function 354 is stored in the storage circuit 34 in the form of a program executable by a computer.

[0066] The transmission function 354 transmits feedback information obtained by providing feedback on the tag candidate information 81 to the medical information processing apparatus. The transmission function 354 is an example of a transmission unit. Specifically, when there is feedback from the user regarding the tag candidate information 81 output from the medical information processing apparatus, the transmission function 354 transmits feedback information indicating the feedback on the input tag candidate information.

[0067] For example, the user refers to the tag candidate information 81 output by the medical information processing apparatus and inputs tag information corresponding to the annotation work for the patient information included in the electronic medical record. During the annotation work, the user may correct the tag candidate information and want the medical information processing apparatus to output the tag candidate information again. Therefore, the transmission function 354 of the processing circuit 350 of the operation terminal 300 according to the first modification transmits feedback information obtained by providing feedback on the tag candidate information to the medical information processing apparatus.

[0068] FIG. 11 is a block diagram showing an example of the functional configuration of a medical information processing apparatus 200 according to a first modification. The processing circuit 230 of the medical information processing apparatus 200 according to the first modification includes, for example, a first acquisition function 231, a second acquisition function 232, a generation function 233, a setting function 234, a determination function 235, an output function 236, a determination function 237, and a third acquisition function 238. In the first modification, each processing function performed by the first acquisition function 231, the second acquisition function 232, the generation function 233, the setting function 234, the determination function 235, the output function 236, the determination function 237, and the third acquisition function 238 is stored in the storage circuit 22 in the form of a program executable by a computer.

[0069] The determination function 237 determines whether there is feedback information. The determination function 237 is an example of a determination unit. Specifically, the determination function 237 determines whether there is feedback information transmitted from the operation terminal 300.

[0070] The third acquisition function 238 acquires feedback information. The third acquisition function 238 is an example of an acquisition unit. Specifically, the third acquisition function 238 acquires the feedback information transmitted from the operation terminal 300. The third acquisition function 238 stores the acquired feedback information in the storage circuit 22.

[0071] Then, the setting function 234 re - sets the instruction sentence information corresponding to the instruction sentence input to the large - language model based on the summary information generated by the generation function 233 and the feedback information acquired by the third acquisition function 238.

[0072] Next, the processing executed by the medical information processing apparatus 200 according to the first modification will be described. FIG. 12 is a flowchart showing an example of the processing content of the medical information processing apparatus 200 according to the first modification. Also, the description of the processing common to the processing executed by the medical information processing apparatus 20 according to the embodiment shown in FIG. 9 is omitted.

[0073] In step S97, the determination function 237 determines whether there is feedback information transmitted from the operation terminal 300 (step S97). Here, if the determination function 237 determines that there is no feedback information transmitted from the operation terminal 300 (step S97: No), the processing of the processing circuit 23 ends. On the other hand, if the determination function 237 determines that there is feedback information transmitted from the operation terminal 300 (step S97: Yes), the process proceeds to step S98.

[0074] In step 98, the setting function 234 re - sets the instruction sentence information corresponding to the instruction sentence to be input to the large - scale language model based on the summary information generated by the generation function 233 and the feedback information acquired by the third acquisition function 238 (step S98).

[0075] The medical information processing apparatus 200 can execute the above - described series of steps while a user such as an operator (for example, a medical staff, etc.) is writing a medical record. Also, when the medical information processing apparatus 200 saves the medical record to the hospital information server 10, tag candidates can be displayed and saved by checking the check box 82.

[0076] As described above, when there is feedback information about the generated tag candidate information 81, the medical information processing apparatus 200 according to the first modification example re - sets the instruction sentence information corresponding to the instruction sentence to be input to the large - scale language model based on the generated summary information and the acquired feedback information. Thereby, the user can provide feedback on the output tag candidate information 81, so that tag candidate information more suitable for the patient is output, and thus more patient - specific annotation work can be performed.

[0077] (Second Modification Example) In the above - described embodiment, the form including the medical history information corresponding to the patient's medical history, the thinking information corresponding to the patient's thinking, and the medical treatment phase information corresponding to the patient's medical treatment phase has been described as the tag candidate, but it is not limited thereto. For example, the tag candidate may include tag candidates including future possibilities.

[0078] A tag candidate including future possibilities refers to a tag candidate that means that, for a patient's current medical history, the condition is speculated to have a not-so-good prognosis (for example, a survival rate of 10%), and the symptoms will change in the future. The change in future symptoms means, for example, changing from the current symptoms to a state such as being cured, maintaining the current state, or deteriorating. Also, as a tag candidate, it may include the thinking information of the family indicating the thoughts and intentions of the patient's family.

[0079] (Third Modification Example) In the above-described embodiment, the medical information processing apparatus 20 acquires text information indicating a plurality of types of medical records at a plurality of time points corresponding to the medical information of the patient, performs natural language processing on the text information, generates summary information 41 corresponding to the text information, and sets command information 61 corresponding to a command sentence to be input to a large language model based on the summary information 41. However, the present invention is not limited to this. The medical information processing apparatus 20 of the third modification example may acquire text information indicating a plurality of types of medical records at a plurality of time points corresponding to the medical information of the patient, and set command information 61 corresponding to a command sentence to be input to a large language model based on the text information.

[0080] For example, the setting function 234 of the medical information processing apparatus 20 sets command information corresponding to a command sentence to be input to a large language model based on the text information of a plurality of types of medical records at a plurality of time points included in the medical information acquired by the second acquisition function 232. Also, since the effects in the third modification example are the same as those in the embodiment, the description thereof is omitted.

[0081] Note that each component of each device illustrated in this embodiment is conceptually functional, and does not necessarily have to be physically configured as shown in the drawings. That is, the specific form of distribution and integration of each device is not limited to that shown in the drawings, and all or part of it can be functionally or physically distributed and integrated in any unit according to various loads, usage situations, etc. Furthermore, each processing function performed by each device can be realized in whole or in any part by a CPU and a program analyzed and executed by the CPU, or can be realized as hardware by wired logic.

[0082] When realizing the technical idea in this embodiment by a medical information processing method, the medical information processing method acquires text information indicating a plurality of types of medical records at a plurality of time points corresponding to a patient's medical information, performs natural language processing on the text information, generates summary information corresponding to the text information, sets instruction information corresponding to an instruction sentence to be input to a large language model based on the summary information, determines tag candidate information indicating tag candidates for attaching to the text information based on the instruction information, and outputs the tag candidate information. Since the processing procedures and effects in the medical information processing method are the same as those in the embodiment, the description is omitted.

[0083] Also, the method described in this embodiment can be realized by executing a pre-prepared program on a computer such as a personal computer or a workstation. This program can be distributed via a network such as the Internet. Also, this program can be recorded on a non-transitory recording medium readable by a computer such as a hard disk, a flexible disk (FD), a CD-ROM, an MO, a DVD, etc., and can also be executed by being read from the recording medium by a computer.

[0084] According to at least one of the embodiments described above, it is possible to provide tag candidates in the annotation work according to the patient's thoughts and medical treatment phases for the medical record.

[0085] Although several embodiments have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, replacements, changes, and combinations of embodiments can be made without departing from the gist of the invention. These embodiments and their modifications are included in the scope and gist of the invention, as well as in the invention described in the claims and its equivalent scope.

Explanation of Reference Numerals

[0086] 1 Medical information system 10 Hospital information server 20 Medical information processing device 30 Operation terminal 21 Communication interface 22 Memory circuit 23 Processing circuit 200 Medical information processing device 231 First acquisition function 232 Second acquisition function 233 Generation function 234 Setting function 235 Decision function 236 Output function 237 Judgment function 238 Third acquisition function 300 Operation terminal

Claims

1. An acquisition unit that acquires text information indicating multiple types of medical records at multiple points in time corresponding to the medical information of a patient; a setting unit that sets command sentence information corresponding to a command sentence to be input to a large-scale language model based on the text information; a determination unit that determines tag candidate information indicating tag candidates to be assigned to the text information based on the command statement information; an output unit that outputs the tag candidate information; A medical information processing device comprising:

2. The acquisition unit acquires patient information of the patient, and acquires the medical information corresponding to the patient information based on the patient information. The medical information processing device according to claim 1 .

3. a generation unit that performs natural language processing on the text information and generates summary information corresponding to the text information; the setting unit sets command sentence information corresponding to a command sentence to be input to a large-scale language model based on the summary information. The medical information processing device according to claim 1 .

4. The medical information is in a format of structured data or unstructured data; The structured data is data whose input format is defined, The unstructured data is data whose input format is not defined. The medical information processing device according to claim 1 .

5. The summary information includes medical history information corresponding to a medical history of the patient, thought information corresponding to a thought of the patient, and medical treatment phase information corresponding to a medical treatment phase of the patient. The medical information processing device according to claim 3 .

6. The tag candidate information is a tag candidate corresponding to the medical history information, the thought information, and the medical treatment phase information. The medical information processing device according to claim 5 .

7. The medical history information includes medical history regarding current illness, past illness, surgical history, and family history; The medical history includes, for each medical history, date and time, disease name, treatment method, and medication information regarding medication; The medical information processing device according to claim 5 .

8. The thought information is information regarding a change in the patient's appearance. The medical information processing device according to claim 5 .

9. The medical treatment phase information includes medical treatment content, The medical care includes disease, recurrence, examination, notification, hospitalization, surgical history, radiation, drug therapy, hormone therapy and genetic testing. The medical information processing device according to claim 5 .

10. the tag candidate information includes the text information, the tag candidates, and check boxes for user selection; The medical information processing device according to claim 6 .

11. The acquisition unit acquires feedback information that has been fed back regarding the tag candidate information, the setting unit resets the command statement information based on the summary information and the feedback information. The medical information processing device according to claim 3 .

12. Obtaining text information indicating multiple types of medical records at multiple points in time corresponding to the patient's medical information; Set command sentence information corresponding to a command sentence to be input to a large-scale language model based on the text information; determining tag candidate information indicating tag candidates to be assigned to the text information based on the command statement information; outputting the tag candidate information; A medical information processing method comprising:

13. On the computer, An acquisition step of acquiring text information indicating multiple types of medical records at multiple points in time corresponding to the medical information of a patient; a setting step of setting command sentence information corresponding to a command sentence to be input to a large-scale language model based on the text information; a determining step of determining tag candidate information indicating tag candidates to be assigned to the text information based on the command sentence information; an output step of outputting the tag candidate information; A medical information processing program for executing the above.

Images