Medical information processing device, medical information processing system, medical information processing method, and program

Abstract

The goal is to enable accurate, comprehensive, and continuous verification of information regarding the subject. [Solution] The medical information processing device of the embodiment comprises an acquisition unit, a calculation unit, a determination unit, and an output control unit. The acquisition unit acquires information regarding the disease risk of a subject. The calculation unit calculates the reliability of the acquired information regarding the disease risk. The determination unit determines, according to the calculated reliability, additional information that is desirable to acquire in addition to the acquired information regarding the disease risk, and the recipient of the request for the additional information. The output control unit outputs request information to request the recipient to provide the determined additional information.

Description

[Technical Field]

[0001] Embodiments disclosed herein and in the drawings relate to medical information processing devices, medical information processing systems, medical information processing methods, and programs. [Background technology]

[0002] Clinical Decision Support Systems (CDS) are known as a support function that helps medical professionals, such as doctors, to correctly understand a patient's condition and make an appropriate diagnosis. CDS have the function of collecting and processing patient information and transmitting it to medical professionals. In this sense, CDS can be considered a medium for facilitating communication (information transmission) within the medical system. A medium is a means of mediation that suppresses variability in semantic interpretation and transmits the meaning of information as accurately as possible. [Prior art documents] [Patent Documents]

[0003] [Patent Document 1] Patent No. 7021101 [Overview of the project] [Problems that the invention aims to solve]

[0004] As a media platform, CDS (Clinical Data Services) needs to be "confirmed" by both patients and healthcare professionals that it is "handling accurate patient information." This can be explained by the characteristic of paradigm (two-way understanding of the meaning of information) that media should possess. However, in emergency situations, such as when the patient is unconscious, patient information may be collected solely by healthcare professionals, potentially leading to insufficient confirmation from the patient. Furthermore, in the midst of a hectic situation, confirmation by healthcare professionals may be overlooked. For example, information that requires confirmation by the patient's family (such as allergies) should also be managed by the patient's family. In addition, the relationship between patients and healthcare professionals, which has been built through regular hospital visits, may introduce bias when transmitting and receiving information. For example, in the case of chronic diseases, it may be difficult to notice the patient's condition changing gradually over time, and the failure to confirm whether accurate patient information is being handled may go unchecked for a long period of time.

[0005] In other words, conventional CDSs only hold patient information obtained at the discretion of healthcare professionals, and in most cases, the accuracy of that information is not verified by the patients themselves. Furthermore, even the accuracy of information verified by healthcare professionals is only partial (there are limits to human perception). Therefore, in order to reduce the risk of misdiagnosis, it is necessary to realize a CDS as a media.

[0006] The problem that the embodiments disclosed herein and in the drawings aim to solve is to enable accurate, complete, and continuous verification of information about a subject. However, the problems that the embodiments disclosed herein and in the drawings aim to solve are not limited to the above problem. Problems corresponding to the effects of each configuration shown in the embodiments described later can also be positioned as other problems. [Means for solving the problem]

[0007] The medical information processing device of this embodiment includes an acquisition unit, a calculation unit, a determination unit, and an output control unit. The acquisition unit acquires information regarding the disease risk of a subject. The calculation unit calculates the reliability of the acquired information regarding the disease risk. The determination unit determines, based on the calculated reliability, additional information that is desirable to acquire in addition to the acquired information regarding the disease risk, and the recipient of the request for the additional information. The output control unit outputs request information to request the recipient to provide the determined additional information. [Brief explanation of the drawing]

[0008] [Figure 1] A diagram showing an example of the configuration of a medical information processing system S including a medical information processing device 1 according to the first embodiment. [Figure 2] A diagram illustrating an example of patient information according to the first embodiment. [Figure 3] A diagram showing an example of the conditions used to calculate the reliability score according to the first embodiment. [Figure 4] A flowchart showing an example of the medical support processing of the medical information processing device 1 according to the first embodiment. [Figure 5] A diagram showing an example of a request screen (for doctors and patients) according to the first embodiment. [Figure 6A] A diagram showing an example of a request screen (for doctors) according to the first embodiment. [Figure 6B] A diagram showing an example of a request screen (for patients) according to the first embodiment. [Figure 7A] A diagram illustrating an example of the input / output of the LLM according to the second embodiment. [Figure 7B] A diagram illustrating another example of the input / output of the LLM according to the second embodiment. [Figure 7C] A diagram illustrating another example of the input / output of the LLM according to the second embodiment. [Figure 8] A diagram illustrating an example of patient information according to the third embodiment. [Modes for carrying out the invention]

[0009] The following describes a medical information processing device, a medical information processing system, a medical information processing method, and a program according to an embodiment, with reference to the drawings. The medical information processing device according to the embodiment provides a means for accurately, completely, and continuously managing information about a subject under the supervision of both the subject and the medical professional. That is, this medical information processing device provides the function of a clinical decision support system (hereinafter also referred to as "CDS-M") as a medium. The CDS-M determines and presents the recipients (subjects, medical professionals, etc.) and the content of additional information that is desirable to collect in order to enhance the reliability of the subject's information. The CDS-M then updates the subject's information based on the collected additional information. Medical professionals include, for example, doctors, nurses, technicians, etc. Subjects include, for example, patients. In the following, the case in which the medical professional is a doctor and the subject is a patient will be used as an example.

[0010] <First Embodiment> [System Configuration] Figure 1 shows an example of the configuration of a medical information processing system S including a medical information processing device 1 according to the first embodiment. The medical information processing system S includes, for example, a medical information processing device 1 that manages information of a target patient (hereinafter referred to as "patient information"), a medical database 3 that stores information that forms the basis of patient information, at least one physician terminal device T1, and at least one patient terminal device T2. These devices and equipment are connected to each other so as to be able to send and receive data via, for example, a communication network NW. The communication network NW includes wireless / wired LANs such as a hospital's backbone LAN (Local Area Network), the Internet network, as well as telephone communication lines, optical fiber communication networks, cable communication networks, and satellite communication networks.

[0011] [Patient Information] In this embodiment, patient information refers to various types of information used in the medical treatment of a patient among all information related to the patient. Patient information is an aggregate of attributes representing an individual patient in any clinical scenario. Attributes include, for example, physical measurement values such as height and weight, vital signs such as blood pressure and pulse, blood test results, presence or absence of diseases, and other content obtained from the patient's answers to the items on the questionnaire (lifestyle, state of motor function, state of cognitive function, state of skin and hair, personality traits, stress and sleep conditions, eating habits, etc.). Patient information may be time-series data. Also, patient information may be a time-series model. A time-series model is data having, for example, the form of "Patient information of Mr. A on May 1, 2024, Patient information of Mr. A on May 4, 2024,...". If a time-series model can be logically and correctly constructed (such as by applying causal inference techniques), problems related to the coherence of CDS-M (realization of sequential connection of communication) can be solved.

[0012] In patient information, a "reliability score" is calculated for each attribute. The reliability score is an index value indicating the certainty of the attribute (certainty of patient information). For example, the reliability score is set such that it increases as the certainty of the attribute increases and decreases as the certainty of the attribute decreases. The medical information processing device 1 basically performs processing to improve this reliability score. The medical information processing device 1 presents patient information to the doctor and / or the patient. At that time, for an attribute with a low reliability score, it requests the doctor and / or the patient to provide additional information. In response, the doctor and / or the patient provide additional information, and the medical information processing device 1 updates the patient information based on the obtained additional information. Patient information is an example of "subject information". The reliability score is an example of "degree of reliability".

[0013] FIG. 2 is a diagram for explaining an example of patient information according to the first embodiment. As shown in FIG. 2, the patient information includes at least one attribute. At least one sub-attribute is associated with each attribute. At least one data item and a reliability score are associated with each sub-attribute. For example, sub-attributes such as "blood glucose" and "lifestyle" are associated with the attribute "diabetes". Further, data items such as value, acquisition date and time, acquisition means, acquisition location, examiner, and evidence are associated with the sub-attributes. Further, a reliability score is associated with the sub-attributes. Note that the sub-attributes may not be set, and data items and a reliability score may be associated with the attributes.

[0014] [Medical information processing device] Returning to FIG. 1, the medical information processing device 1 controls the overall operation of the medical information processing system S and manages patient information. The medical information processing device 1 is an example of a "medical information processing device" or a "medical information processing system". The medical information processing device 1 may be, for example, a workstation, a server, or the like. The medical information processing device 1 includes, for example, a processing circuit 10, a communication interface 20, and a memory 30. The communication interface 20 communicates with external devices such as a medical database 3, a doctor terminal device T1, and a patient terminal device T2 via a communication network NW. The communication interface 20 includes, for example, a communication interface such as a NIC (Network Interface Card).

[0015] The processing circuit 10 controls the overall operation of the medical information processing device 1. The processing circuit 10 includes, for example, an acquisition function 11, a calculation function 12, an extraction function 13, a determination function 14, an output control function 15, and a management function 16. The processing circuit 10 realizes these functions by, for example, a hardware processor (computer) executing a program stored in the memory 30 (storage circuit).

[0016] A hardware processor refers to circuits such as a CPU (Central Processing Unit), GPU (Graphics Processing Unit), Application Specific Integrated Circuit (ASIC), and programmable logic devices (e.g., Simple Programmable Logic Device (SPLD) or Complex Programmable Logic Device (CPLD), Field Programmable Gate Array (FPGA)). Instead of storing the program in memory 30, the hardware processor may be configured to directly embed the program within its circuitry. In this case, the hardware processor functions by reading and executing the program embedded within its circuitry.

[0017] The above program may be stored in memory 30 in advance, or it may be stored in a non-temporary storage medium such as a DVD or CD-ROM, and installed from the non-temporary storage medium to memory 30 when the non-temporary storage medium is mounted in the drive device (not shown) of the medical information processing device 1. The hardware processor is not limited to being configured as a single circuit, but may be configured as a single hardware processor by combining multiple independent circuits to realize each function. Alternatively, multiple components may be integrated into a single hardware processor to realize each function.

[0018] The acquisition function 11 acquires various information ("information regarding the disease risk of the subject") from external devices via a communication network NW. The acquisition function 11 acquires, for example, information that forms the basis of patient information (hereinafter also referred to as "original patient information") from the medical database 3. Original patient information includes, for example, various test data of the patient (measured values, images, etc.), electronic medical records, and response data to questionnaires. Alternatively, the acquisition function 11 may acquire test information from other various medical devices. The acquisition function 11 is an example of an "acquisition unit".

[0019] The calculation function 12 calculates the Credibility Score of Information (CSI) of the information acquired by the acquisition function 11. The calculation function 12 calculates the Credibility Score from the perspective of the 5Ws (who, what, where, when, why). The calculation function 12 calculates the Credibility Score based on the following formula (1), for example.

[0020] CSI=f who +f what +f where +f when +f why ...(1)

[0021] In equation (1) above, f is the decision function. The decision function f outputs, for example, "0" if the condition is met, and "-1" if the condition is not met. The decision function f for each perspective is defined based on the conditions shown in Figure 3. For example, the decision function f for the "who" perspective who One of the sub-perspectives is "approver," and the condition for this is "whether the results of a certain test are registered by a person with the appropriate authority." who The system outputs "0" if the patient information used to calculate the reliability score meets this condition, and "-1" if it does not. For example, in the patient information example shown in Figure 2, the condition is determined by comparing the data item "Test Performer" with pre-set reference data according to the attribute type.

[0022] Also, for example, f, which is a determination function from the perspective of "when", when has "acquisition date and time" set as a sub - perspective, and as its condition, "Is the information acquisition time not too far from the current time?" is set. f when outputs "0" when the patient information to be calculated for the reliability score meets this condition, and outputs "-1" when it does not meet this condition. For example, in the example of patient information shown in Figure 2, by comparing the data of the data item "acquisition date and time" with the current time, it is determined whether the condition is met.

[0023] The calculation function 12 determines that the reliability score is "Low CSI" if the CSI calculated using, for example, the above formula (1) is less than 0, and determines that the reliability score is "High CSI" if the CSI is 0, and registers the determination result in the patient information. Incidentally, the calculation function 12 may register the numerical value of the CSI calculated using, for example, the above formula (1) in the patient information.

[0024] Incidentally, it is not necessary to use all of the 5W perspectives in the above formula (1), and a part (at least one) of the 5W perspectives may be used. Also, each perspective (for example, why) may be defined separately into a doctor's perspective and a patient's perspective. A perspective of "reputation", which is different from the 5W, may be set. The determination function f of this reputation perspective may be defined to perform determination processing based on conditions such as "Is it based on authoritative information?" and "Is there a second opinion?".

[0025] The calculation function 12 is an example of a "calculation unit". The reliability score is an example of "reliability". That is, the calculation function 12 calculates the reliability score of information regarding disease risk.

[0026] Returning to Figure 1, the extraction function 13 extracts a predetermined number of patient information entries with low reliability scores when the patient information is sorted in order of reliability score magnitude. For example, the extraction function 13 sorts the attributes that have been determined to have a "low reliability score (Low CSI)" in order of reliability score magnitude, and extracts a predetermined number of attributes (for example, the three worst ones) in descending order of reliability score. The extraction function 13 is an example of an "extraction unit".

[0027] The judgment function 14 determines, based on a predetermined number of patient information extracted by the extraction function 13 (according to the reliability score calculated by the calculation function 12), what additional information is desirable to acquire in addition to the information acquired by the acquisition function 11, and where to request this additional information. For example, the judgment function 14 determines information regarding aspects of the patient information for which data is missing, as additional information. For example, the judgment function f for the "when" aspect. when If the result is "-1", the judgment function 14 determines that additional information should be obtained from the perspective of "when". Also, if the reliability score of the attribute "disease name (e.g., cancer)" is low, the judgment function 14 determines that the request should be made to a doctor, not the patient. Also, if the reliability score of the attribute "information that is not to be made public to the patient (e.g., internal approval history)" is low, the judgment function 14 determines that the request should be made to a doctor, not the patient. Note that the request may include the patient's cohabitants or caregivers (family members, etc.) instead of the patient. The judgment function 14 is an example of a "judgment unit". The judgment function 14 determines, according to the reliability score, additional information that is desirable to obtain in addition to the disease risk information, and the recipient of the request for the additional information.

[0028] The output control function 15 outputs request information to request additional information from the requesting party. The output control function 15 outputs a screen to the terminal device (doctor terminal device T1 and / or patient terminal device T2, etc.) to request additional information from the requesting party (doctor and / or patient, etc.). This screen includes information (attributes) related to disease risk, a reliability score, and string information requesting additional information from the requesting party. For example, if the judgment function 14 determines that information from the perspective of "when" is to be used as additional information and the requesting party is determined to be a patient, the output control function 15 outputs information to the patient terminal device T2 to display a screen containing the message, "Please tell me when you fell." The output control function 15 may also output request information as a text message or voice (machine voice) instead of (or in addition to) a screen. The output control function 15 is an example of an "output control unit."

[0029] The management function 16 manages patient information PI stored in memory 30. The management function 16 associates the reliability score calculated by the calculation function 12 with attributes and stores it in memory 30 as patient information PI. The management function 16 also adds and deletes attributes included in the patient information. For example, the management function 16 adds and deletes attributes and sub-attributes using information represented in trees / graphs, such as medical ontology and disease concept networks. Furthermore, if the electronic medical record created by the physician contains a description of "suspected diabetes" (a description of the disease name), the management function 16 adds attributes (sub-attributes) related to "diabetes." Also, if the electronic medical record created by the physician contains a description of "having seizures" (a description of symptoms), the management function 16 adds the underlying disease (cerebrovascular disease, metabolic disorder, etc.) and information that needs to be gathered (age of initial onset, medical history, family history, drug administration history, etc.) as attributes (sub-attributes). Alternatively, if the patient's completed medical questionnaire includes a note saying "I might have a fracture" (indicating a disease), the management function 16 adds an attribute (sub-attribute) related to "fracture." Furthermore, if the patient's completed medical questionnaire includes a note saying "My leg hurts" (indicating a symptom), the management function 16 adds the causative disease (fracture, muscle strain, etc.) and other information requiring further inquiry (onset date, etc.) as attributes (sub-attributes). In addition, the management function 16 may update the patient information attributes and the values ​​of each data item based on information in the hospital system, such as information obtained through the medical questionnaire and entered into the physician's terminal device T1. The management function 16 may also add attributes from guidelines or standard protocols (e.g., emergency scenarios).

[0030] Management function 16 is an example of a "management unit." Specifically, management function 16 stores patient information based on disease risk and reliability scores in a memory device.

[0031] Memory 30 can be implemented by semiconductor memory elements such as RAM (Random Access Memory) or flash memory, a hard disk, or an optical disc. These non-transient storage media may also be implemented by other storage devices connected via a communication network NW, such as a NAS (Network Attached Storage) or an external storage server device. Memory 30 may also include non-transient storage media such as ROM (Read Only Memory) or registers. Memory 30 stores, for example, patient information PI. In addition, memory 30 stores programs, parameter data, and other data used by the processing circuit 10. Memory 30 is an example of a "storage device".

[0032] [Medical Database] Medical Database 3 is a storage device that stores original patient information, which forms the basis of patient information. Medical Database 3 stores, for example, various patient test data (measurements, images, etc.), electronic medical records, and response data to questionnaires. Medical Database 3 is implemented using, for example, semiconductor memory elements such as RAM and flash memory, hard disks, and optical discs.

[0033] [Physician terminal device] The physician terminal device T1 is a device for accessing various information provided by the medical information processing device 1. The physician terminal device T1 is operated, for example, by physician D. The physician terminal device T1 is, for example, a personal computer, a tablet, a smartphone, or other mobile terminal. The physician terminal device T1 includes, for example, a communication function for data communication with other devices, an input interface function for receiving various instructions from physician D, and a display function for displaying various information.

[0034] [Patient terminal device] The patient terminal device T2 is a device for accessing various information provided by the medical information processing device 1. The patient terminal device T2 is operated, for example, by patient P. The patient terminal device T2 is, for example, a personal computer, a tablet, a smartphone or other mobile device. The patient terminal device T2 is equipped with, for example, a communication function for data communication with other devices, an input interface function for receiving various instructions from patient P, and a display function for displaying various information.

[0035] [Processing flow] Next, we will explain the sequence of medical support processing in the medical information processing device 1. Figure 4 is a flowchart showing an example of medical support processing in the medical information processing device 1 according to the first embodiment. The medical support processing shown in Figure 4 is started, for example, when a physician D who is treating patient P issues an instruction to start medical support processing via the input interface of the physician terminal device T1. In this case, patient P and physician D may be in the same space, such as a hospital examination room (face-to-face consultation), or in separate spaces (online consultation).

[0036] First, the acquisition function 11 acquires original patient information, which is the basis for patient information, from the medical database 3 via the communication network NW (step S101). Original patient information includes, for example, various test data of the patient (measured values, images, etc.), electronic medical records, and response data to questionnaires. For example, the acquisition function 11 acquires original patient information from the medical database 3 using the patient identifier (patient ID) that identifies the patient, entered by physician D operating the physician terminal device T1, as the key. The acquisition function 11 registers the acquired original patient information in the patient information PI according to its attributes. Furthermore, if past patient information of patient P is stored in memory 30, the acquisition function 11 also acquires this past patient information.

[0037] Next, the calculation function 12 calculates the reliability score of the original patient information obtained by the acquisition function 11 (step S103). The calculation function 12 calculates the reliability score, for example, from the perspective of the 5Ws (who, what, where, when, why) and registers it in the patient information PI.

[0038] Next, the extraction function 13 sorts each attribute included in the patient information PI in order of the magnitude of its reliability score and extracts a predetermined number of attributes (patient information) with low reliability scores (step S105). For example, the extraction function 13 sorts the attributes that have been determined to have a "low reliability score (Low CSI)" in order of the magnitude of their reliability scores and extracts a predetermined number of attributes (for example, the three worst ones) in descending order of reliability score.

[0039] Next, the determination function 14 determines, with respect to the attributes extracted by the extraction function 13, what additional information is desirable to acquire in addition to the information acquired by the acquisition function 11, and the destination for requesting this additional information (step S107).

[0040] The output control function 15 outputs request information to request the recipient to provide additional information (step S109). For example, the output control function 15 outputs a screen to the terminal device (doctor terminal device T1 and / or patient terminal device T2, etc.) to request the recipient (doctor and / or patient, etc.) to provide additional information. As a result, the doctor terminal device T1 and / or patient terminal device T2 display a screen for requesting the provision of additional information (hereinafter referred to as the "request screen").

[0041] Figure 5 shows an example of a request screen (for doctors and patients) according to the first embodiment. The request screen in Figure 5 displays information about four attributes (A, B, C, D) with low reliability scores. On this request screen, for example, for attribute A, in order to improve the reliability score from the perspective of "when," the message "To: Patient When did the symptoms of XX appear?" is displayed as a confirmation point requesting additional information from the patient to whom the request is being made. Also, on this request screen, for example, for attribute B, in order to improve the reliability score from the perspective of "who," the message "To: Doctor Please confirm approval of △△" is displayed as a confirmation point requesting additional information from the doctor to whom the request is being made. After viewing such a request screen, doctors and / or patients can operate the doctor terminal device T1 and / or the patient terminal device T2 to input their answers to the confirmation points.

[0042] Figure 6A shows an example of a request screen (for physicians) according to the first embodiment. In the request screen of Figure 6A, information is displayed for two attributes (B and C) that have low reliability scores and for which the request is directed to a physician. In this request screen, for example, for attribute C, the message "To: Physician, γ-GTP is too high." is displayed as a confirmation point requesting additional information from the physician to improve the reliability score in terms of "what". A physician who has seen such a request screen can operate the physician terminal device T1 to input their response to the confirmation point.

[0043] Figure 6B shows an example of a request screen (for patients) according to the first embodiment. In the request screen of Figure 6B, information is displayed for three attributes (A, C, D) that have low reliability scores and for which the request is directed to a patient. In this request screen, for example, for attribute C, in order to improve the reliability score from the perspective of "why," the message "To: Patient, did you drink excessively last night?" is displayed as a confirmation point requesting additional information from the patient. A patient who sees such a request screen can operate the patient terminal device T2 to input their answers to the confirmation points. Such a patient-oriented request screen may also be presented to the patient (patient terminal device T2) when filling out a medical questionnaire in the pre-consultation stage. Furthermore, compared to the request screen for doctors, the wording used in the patient-oriented request screen may use simple language without using technical terms.

[0044] Returning to Figure 4, the acquisition function 11 then acquires additional information from the physician terminal device T1 and / or the patient terminal device T2, and the management function 16 updates the patient information PI stored in memory 30 using the acquired additional information (step S111). Furthermore, the calculation function 12 recalculates the reliability score using the patient information to be updated (information on disease risk) and the additional information, and the management function 16 updates the patient information PI stored in memory 30 using the recalculated reliability score. Here, the management function 16 may also add or delete attributes themselves in the patient information PI stored in memory 30.

[0045] Next, if physician D operates the physician terminal device T1 to input a treatment termination instruction, and the acquisition function 11 accepts this treatment termination instruction (step S113; YES), the processing of this flowchart ends. On the other hand, if the acquisition function 11 does not accept the treatment termination instruction (step S113; NO) (for example, if the acquisition function 11 accepts an instruction to recalculate the reliability score), the process returns to step S103 and the subsequent processing is repeated. As this process is repeated, it is expected that towards the end of the treatment, the attributes necessary for the treatment will remain, and the reliability score of each attribute will be high.

[0046] As described above, the first embodiment enables accurate, complete, and continuous verification of patient information by both the subject and healthcare professionals. This allows for the systematic organization and visualization of patient information managed in the Clinical Decision Support System (CDS). Furthermore, it reduces the risk of misdiagnosis by the Clinical Decision Support System. Moreover, by using such patient information, highly reliable input data can be provided to other applications such as various diagnostic support AI (Artificial Intelligence) models. In addition, it becomes possible to handle patient information more accurately when the subject and healthcare professionals are spatially and temporally separated, such as in telemedicine.

[0047] <Second Embodiment> The second embodiment will be described below. In the following description, components and functions identical to those in the first embodiment will be denoted by the same reference numerals, and detailed explanations will be omitted. The second embodiment differs from the first embodiment in that the patient information to be managed is a collection of parameters (weight parameters). The parameters are, for example, the weight parameters of a Large Language Model (LLM). That is, the patient information includes patient information (information on disease risk) and the weight parameters of a large language model that are set by learning reliability scores.

[0048] In the second embodiment, the management function 16 inputs the original patient information acquired by the acquisition function 11 into the LLM. This allows the LLM to manage the patient information (using its own weight parameters). Figure 7A is a diagram illustrating an example of the input and output of the LLM in the second embodiment. The management function 16 inputs the original patient information acquired by the acquisition function 11 and a first prompt to output the patient information contained in the LLM into the LLM. This causes the LLM to output the patient information contained in the LLM in a table format consisting of a set of attributes, as shown in Figure 2 of the first embodiment, in a format that is easy for humans to read.

[0049] Figure 7B illustrates another example of the input and output of the LLM according to the second embodiment. The calculation function 12 inputs the original patient information acquired by the acquisition function 11 and a second prompt for calculating a reliability score for the original patient information into the LLM. This second prompt includes an instruction to calculate the reliability score using the same calculation logic (5W) as in the first embodiment. As a result, the LLM outputs the patient information (with reliability score) contained in the LLM in a table format consisting of a set of attributes as shown in Figure 2 of the first embodiment, in a format that is easy for humans to read.

[0050] Figure 7C illustrates another example of input and output of the LLM according to the second embodiment. The extraction function 13 and the determination function 14 input the patient ID and a third prompt for calculating additional information and the request destination to the LLM. This third prompt includes instructions for performing functions similar to those of the extraction function 13 and the determination function 14 in the first embodiment. As a result, the LLM outputs the additional information and the request destination (e.g., information indicating the format for receiving and responding to questions from the request destination).

[0051] The above LLM input / output is just one example, and various configurations are possible. For example, the first to third prompts shown above, which are an example of input data, may be combined into a single prompt.

[0052] As described above, the second embodiment enables accurate, complete, and continuous verification of patient information by both the subject and healthcare professionals. Furthermore, by managing weight parameters as subject information (patient information), subject information with diverse formats and characteristics can be managed comprehensively and simply. Additionally, by rewriting the prompts entered into the LLM, it becomes possible to provide subject information in a desired format depending on the application, thereby expanding its range of uses.

[0053] <Third Embodiment> The third embodiment will be described below. In the following description, components and functions identical to those in the first embodiment will be denoted by the same reference numerals as in the first embodiment, and detailed explanations will be omitted. The third embodiment differs from the first embodiment in that the patient information to be managed is a simulation model. That is, the patient information includes a simulation model generated based on patient information (information on disease risk).

[0054] Figure 8 illustrates an example of patient information according to the third embodiment. Figure 8 shows a case where the patient information is a cardiac simulation model, and the attributes managed are simulationable data related to the heart (e.g., left ventricular ejection fraction (LVEF)). When calculating the reliability score, the calculation function 12 performs calculations based on the simulation results in addition to the same calculation logic (5W) as in the first embodiment. The calculation function 12 compares the actual data (e.g., measured LVEF) with the simulation results of the simulation model (e.g., predicted LVEF), and if the simulation result (predicted value) is far removed from the actual data (measured value), or if it could not be calculated at all, it calculates a low reliability score for that data. In this case, the judgment function 14 determines whether to provide additional information to the requesting physician D, such as "confirmation of measured value (request for remeasurement)" or "approval request to update the simulation model so that the output of the simulation model approaches the measured value." In response, the output control function 15 outputs request information to the physician terminal device T1 to request the recipient to provide additional information.

[0055] According to the third embodiment described above, both the subject and the medical professional can continuously and accurately verify the subject's information without any omissions. Furthermore, by managing the simulation model as the subject's information (patient information), calculating its reliability, and providing suggestions for improvement, the accuracy of the simulation model can be improved.

[0056] Furthermore, some or all of the configurations exemplified in the first to third embodiments described above may be implemented in combination.

[0057] While several embodiments have been described, these embodiments are presented as examples only and are not intended to limit the scope of the invention. These embodiments can be carried out in a variety of other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their variations are included in the scope and spirit of the invention, as well as in the claims and their equivalents. [Explanation of Symbols]

[0058] 1. Medical Information Processing Device 3. Medical Databases 10 Processing Circuit 11. Acquisition function 12. Calculation Function 13. Extraction function 14. Judgment Function 15. Output control function 16 Management functions 20 Communication Interfaces 30 memory NW (Network Communication Network) S Medical Information Processing System T1 Physician Terminal Device T2 Patient Terminal Device

Claims

1. An acquisition unit that acquires information on the disease risk of the subject, A calculation unit that calculates the reliability of the acquired disease risk information, A determination unit determines, based on the calculated confidence level, what additional information is desirable to acquire in addition to the acquired disease risk information, and where to request the additional information. An output control unit outputs request information to request the requesting party to provide the determined additional information, A medical information processing device equipped with [a specific feature].

2. The output control unit outputs a screen to the terminal device for requesting the provision of the additional information from the requesting party. The medical information processing device according to claim 1.

3. The output control unit outputs a screen to the terminal device of the medical professional requesting the medical professional to provide the additional information. The medical information processing device according to claim 1.

4. The output control unit outputs a screen to the terminal device of the subject, which is the requesting party, requesting the provision of the additional information. The medical information processing device according to claim 1.

5. The screen includes information regarding the disease risk, the confidence level, and a string of characters requesting the recipient to provide the additional information. A medical information processing device according to any one of claims 2 to 4.

6. The system further includes a management unit that stores information regarding the disease risk and subject information based on the confidence level in a storage device. A medical information processing device according to any one of claims 1 to 4.

7. The acquisition unit acquires the additional information provided by the requesting party in accordance with the request information. The management unit updates the subject information stored in the storage device using the acquired additional information. The medical information processing device according to claim 6.

8. The calculation unit recalculates the confidence level using the disease risk information and the additional information. The management unit updates the subject information stored in the storage device using the recalculated confidence level. The medical information processing device according to claim 7.

9. The system further includes an extraction unit that extracts a predetermined number of subject information with low reliability when the subject information is arranged in order of the magnitude of the reliability, The determination unit determines the additional information and the request destination for the extracted predetermined number of subject information. The medical information processing device according to claim 6.

10. The subject information includes information on disease risk and weight parameters of a large-scale language model set by learning the confidence level. The medical information processing device according to claim 6.

11. The subject information includes a simulation model generated based on the disease risk information, The medical information processing device according to claim 6.

12. A medical information processing device according to any one of claims 1 to 4, A storage device that stores subject information linked to the aforementioned reliability level and information regarding the disease risk, A medical information processing system equipped with [a specific feature / feature].

13. Computers Obtain information on the disease risk of the subject, The reliability of the acquired information regarding the disease risk is calculated, Based on the calculated confidence level, the system determines what additional information is desirable to obtain in addition to the disease risk information obtained, and where to request such additional information. Output request information to request the requesting party to provide the determined additional information. Medical information processing method.

14. On the computer, Obtain information about the disease risk of the subject, The reliability of the acquired information regarding disease risk is calculated. Based on the calculated confidence level, the system determines what additional information is desirable to obtain in addition to the disease risk information obtained, and where to request such additional information. The system outputs request information to request the recipient to provide the determined additional information. program.

Images