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

Abstract

To efficiently support determination of a treatment context of an individual subject.SOLUTION: A medical information processing apparatus includes a path specification unit and a presentation unit. The processing circuitry specifies a path that connects a node representing a first state of a subject and a node representing a second state of the subject and defines a relationship between the nodes, based on a plurality of graphs in which a relationship between the first state, a clinical task representing a medical practice performed for the first state, and the second state of the subject that changes by performing the clinical task is defined by nodes, and a time series of the clinical tasks. The presentation unit presents context information representing a relationship between nodes connected by a path.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

[0002] Traditionally, in intensive care units (ICUs) at hospitals and other medical facilities, a single patient is cared for 24 hours a day by a team of multidisciplinary staff. When providing such care, the person in charge of patient care collects information on the patient's condition before they took over their care and information on the patient during their absence, and then decides on a course of treatment and procedures.

[0003] However, such information is generally scattered across various departmental systems. It is also difficult to know who recorded what information and when, and collecting the information is time-consuming. For example, if an order is changed, the person in charge cannot easily determine who made the change and why, which can make it take a long time to decide on a course of action.

[0004] For these reasons, there is a need to efficiently determine the treatment context for each patient. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Patent No. 6780520 Summary of the Invention [Problem 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 support efficient determination of the treatment context for an individual subject. However, the problems to be solved by the embodiments disclosed in this specification and the drawings are not limited to the above problem. Problems corresponding to the effects of each configuration shown in the embodiments described below can also be positioned as other problems. [Means for solving the problem]

[0007] A medical image processing apparatus according to an embodiment includes a path identification unit and a presentation unit. The path identification unit identifies paths that connect nodes representing the first state and nodes representing the second state and that define the relationships between the nodes, based on a plurality of graphs in which the relationships between a first state of a subject, clinical tasks representing medical procedures performed for the first state, and a second state of the subject that changes as a result of performing the clinical tasks, and a time series of the clinical tasks. The presentation unit presents context information that indicates the relationships between the nodes connected by the paths. [Brief explanation of the drawings]

[0008] [Figure 1] FIG. 1 is a diagram illustrating an example of the configuration of a medical information processing system according to an embodiment. [Figure 2] FIG. 2 is a diagram illustrating an example of a clinical card according to the embodiment. [Figure 3] FIG. 3 is a diagram illustrating an example of a path identification process according to the embodiment. [Figure 4] FIG. 4 is a diagram illustrating an example of a path identification process according to the embodiment. [Figure 5] FIG. 5 is a diagram illustrating an example of the operation of each function of the medical image processing apparatus according to the embodiment. [Figure 6] FIG. 6 is a diagram illustrating an example of utilization of context information according to the embodiment. [Figure 7] FIG. 7 is a diagram illustrating an example of utilization of context information according to the embodiment. [Figure 8]FIG. 8 is a flowchart showing an example of processing executed by the medical image processing apparatus according to the embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0009] Hereinafter, embodiments of a medical information processing apparatus, a medical information processing system, and a medical information processing method will be described in detail with reference to the drawings.

[0010] FIG. 1 is a diagram showing an example of the configuration of a medical information processing system according to this embodiment.

[0011] The medical information processing system according to this embodiment is installed in a medical facility such as a hospital and supports efficient determination of the treatment context for an individual subject. Specifically, the medical information processing system according to this embodiment constructs multiple general-purpose graphs that define the relationships between nodes, each representing a subject's condition or a medical procedure performed on the subject. Furthermore, the medical information processing system identifies paths connecting the nodes based on the subject's condition or the results of the medical procedure performed on the subject, thereby presenting subject-specific context information that represents the relationships between the nodes.

[0012] 1, a medical information processing system 100 according to this embodiment includes an information integration system 110 and a medical information processing device 120. The information integration system 110 and the medical information processing device 120 are connected to each other via a network 130 such as a LAN (Local Area Network) or a WAN (Wide Area Network) so as to be able to communicate with each other.

[0013] The information integration system 110 includes a patient information DB (Data Base) 111. For example, the information integration system 110 is realized by computer equipment such as a server or a workstation equipped with an electronic medical record system, a HIS (Hospital Information System), a RIS (Radiology Information System), or the like.

[0014] The patient information DB 111 stores various types of medical information about patients. A patient is an example of a subject. For example, the patient information DB 111 stores, as medical information for multiple patients, information about each patient's illness (pre-existing conditions, complications, side effects, etc.), the results of various measurements and tests for each patient, symptoms (pathological conditions) for treatments (including administered drugs), connected devices, etc.

[0015] The medical information processing device 120 is a device used by users who are medical professionals involved in multidisciplinary rounds (e.g., attending physicians, ICU doctors, ICU nurses, pharmacists, physical therapists, clinical engineers, etc.), and has a function for supporting decision-making on treatment strategies for patients admitted to the ICU. For example, the medical information processing device 120 is realized by a personal computer or the like.

[0016] Specifically, the medical information processing device 120 includes a network (NW) interface 121, a storage circuitry 122, an input interface 123, a display 124, a camera 125, a microphone 126, and a processing circuitry 127.

[0017] The NW interface 121 controls the transmission and communication of various data transmitted and received between the medical information processing device 120 and other devices connected via the network 130. Specifically, the NW interface 121 is connected to the processing circuitry 127, and outputs patient information received from the information integration system 110 to the processing circuitry 127. For example, the NW interface 121 is realized by a network card, a network adapter, a NIC (Network Interface Controller), or the like.

[0018] The storage circuitry 122 stores various data, various programs, etc. For example, the storage circuitry 122 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.

[0019] Specifically, the storage circuit 122 includes a clinical card DB 122a and a context DB 122b.

[0020] The clinical card DB 122a stores a plurality of clinical cards. For example, a clinical card is a graph in which nodes represent the current state (symptoms, etc.) of a generalized patient, not an individual, and clinical tasks (examinations, treatments, etc.) representing medical procedures for the patient, and the plurality of nodes are arranged based on the relationships between the nodes. The clinical card DB 122a stores information representing the patient's state or medical procedures and the clinical card in association with each other.

[0021] Clinical cards are constructed by arranging nodes related to specific medical procedures or specific diseases. Although the scope of a single medical card is limited and it is not possible to represent the entire condition of a specific subject, a large network of nodes can be formed by connecting clinical cards based on the condition of an individual patient, the medical procedures to be performed on that individual patient, or the medical procedure plan, using the processing described below.

[0022] This allows for the construction of a large-scale node network that is useful for efficiently and accurately predicting the condition of individual patients and their future conditions.

[0023] The context DB 122b stores context information for each patient. For example, the context information is a graph that represents the relationship between the patient's condition and the clinical tasks performed for that condition using paths connecting nodes. The context DB 122b stores a patient ID that identifies the patient and the context information in association with each other.

[0024] The input interface 123 receives input operations of various instructions and various information from a user. Specifically, the input interface 123 is connected to the processing circuitry 127, converts the input operations received from the operator into electrical signals, and outputs the electrical signals to the processing circuitry 127.

[0025] For example, the input interface 123 is realized by a trackball for performing various inputs, a switch button, a mouse, a keyboard, a touchpad for performing input operations by touching the operation surface, a touchscreen in which the display screen and touchpad are integrated, a non-contact input circuit using an optical sensor, and a voice input circuit.

[0026] In this specification, the input interface 123 is not limited to an interface having physical operation parts such as a mouse, a keyboard, etc. For example, an example of the input interface 123 also includes 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 a control circuit.

[0027] The display 124 displays various types of information and various types of data. Specifically, the display 124 is connected to the processing circuit 127 and displays various types of information and various types of data output from the processing circuit 127. For example, the display 124 is realized by a liquid crystal monitor, a CRT (Cathode Ray Tube) monitor, a touch panel, or the like.

[0028] The camera 125 captures an image of a medical professional or a patient. Note that a plurality of cameras 125 may be installed within a medical facility.

[0029] The microphone 126 records the voices of the medical staff or the patient. Note that multiple microphones 126 may be installed within a medical facility.

[0030] The processing circuitry 127 controls the components of the medical image processing device 120 in response to input operations received from the operator via the input interface 123 .

[0031] Specifically, the processing circuit 127 has an acquisition function 127a, a card identification function 127b, a collection function 127c, a path identification function 127d, and a presentation function 127e. Here, the acquisition function 127a is an example of an acquisition unit. The card identification function 127b is an example of a graph identification unit. The path identification function 127d is an example of a path identification unit and a node identification unit. The presentation function 127e is an example of a presentation unit.

[0032] The acquisition function 127a acquires medical data indicating the current condition of a target patient to be processed or a medical procedure to be performed on the target patient. For example, the acquisition function 127a accepts input of information indicating the current condition of the target patient or a medical procedure to be performed on the target patient from a user via the input interface 123. The current condition of the patient may be information expressing the patient's symptoms or diagnosis, or may be information indicating the test results of various tests.

[0033] Alternatively, information obtained by processing the test results of various tests using a determination algorithm or the like may be stored. For example, whether or not a numerical value obtained from a blood test is abnormal can be determined by applying threshold processing to the numerical value. Instead of receiving the blood test numerical value itself, information indicating whether or not a specific blood test item is abnormal, obtained as a result of threshold processing, may be received. Also, these pieces of information may be combined.

[0034] Specifically, when a target patient transported to an ICU is found to have symptoms of respiratory failure, the attending physician, who is a user, inputs information indicating "respiratory failure" as the target patient's condition. The acquisition function 127a acquires this information as medical data indicating the target patient's condition. Furthermore, an ICU doctor, who is another user, inputs information indicating "artificial respiration" as a planned medical procedure for "respiratory failure." The acquisition function 127a acquires this information as medical data indicating a medical procedure for "respiratory failure."

[0035] Similarly, if symptoms of renal failure are observed, the attending physician inputs information indicating "renal failure" as the condition of the target patient. The acquisition function 127a acquires this information as medical data indicating the condition of the target patient. Furthermore, the doctor providing treatment to the target patient inputs information indicating "dialysis" as the medical procedure planned for "renal failure." The acquisition function 127a acquires this information as medical data indicating the medical procedure for "renal failure."

[0036] In the following example, a target patient is admitted to an ICU to receive treatment, but the admission destination is not limited to this and may be a department where treatment is received or a specialized medical institution. In this embodiment, when the target patient is admitted to the ICU, the acquisition function 127a acquires information indicating the target patient's condition at the time of admission to the ICU as medical data. In addition, the acquisition function 127a acquires medical data every time a medical procedure is performed by a medical professional.

[0037] The card identification function 127b identifies a clinical card having nodes related to clinical tasks that represent the condition of a patient or medical procedures to be performed on a patient indicated in the medical data acquired by the acquisition function 127a. A clinical card is an example of a graph.

[0038] For example, the card identification function 127b refers to the clinical card DB 122a in the storage circuit 122 and identifies the clinical card corresponding to the input content received by the acquisition function 127a.

[0039] Specifically, if the input content received by the acquisition function 127a is information representing the patient's condition of "respiratory failure," the card identification function 127b identifies "respiratory failure," which is the clinical card corresponding to "respiratory failure." Also, if the input content received by the acquisition function 127a is information representing the clinical task of "artificial respiration," the card identification function 127b identifies "artificial ventilator," which is the clinical card corresponding to "artificial respiration."

[0040] Similarly, if the input content received by the acquisition function 127a is information representing the patient's condition of "renal failure," the card identification function 127b identifies "renal failure," which is the clinical card corresponding to "renal failure." Furthermore, if the input content received by the acquisition function 127a is information representing the clinical task of "dialysis," the card identification function 127b identifies "dialysis," which is the clinical card corresponding to "dialysis."

[0041] Here, Fig. 2 is a diagram for explaining an example of a clinical card, which shows an example of a clinical card CC1 for "dialysis" and a clinical card CC2 for "artificial respirator."

[0042] Clinical card CC1 has a node n11 for “normal fluid volume,” a node n12 for “low blood pressure,” a node n13 for “water removal discontinued,” a node n14 for “dyspnea,” and a node n15 for “increased fluid volume.” Note that, like node n11 and node n15, nodes of the same type but with different time series (in this example, the contents of the nodes are common in that they are information related to fluid volume) may be arranged.

[0043] Specifically, the test value at the time of transport is input into the "normal hydration" node n11, and another node may be provided to input the normal value of "hydration" and the target value after treatment. The input of information into each node provided on the clinical card described above may simply be the input of the type of planned medical treatment, or may also include the input of details of the medical treatment.

[0044] For example, when entering a medical procedure such as "artificial respiration," parameters and additional information detailing the content of the medical procedure, such as the type of artificial respiration equipment to be used, the planned length of the medical procedure, or the amount of oxygen to be administered, can be entered into each node.

[0045] Furthermore, the clinical card specifies the chronological order between nodes that can be connected. In the clinical card CC1 in Figure 2, the chronological order between nodes is specified in advance according to the patient condition, medical intervention, and their causal relationships, such as "node n11 → node n13 → node n15" or "node n12 → node n13 → node n14."

[0046] Furthermore, the nodes in FIG. 2 are displayed as unspecified nodes that are not adopted as context information, which will be described later, but nodes that are adopted as context information are displayed as specified nodes.

[0047] 2 are all displayed as unspecified paths, but paths specified by a path specifying function 127d (described later) are displayed as specified paths.

[0048] Additionally, the clinical card CC1 contains related information r12 representing "vital signs" associated with node n12, related information r13 representing "dialysis setting values" associated with node n13, related information r14 representing "observation items" associated with node n14, and related information r15 representing "vital signs" associated with node n15. The path identification function 127d identifies a path representing the context of the target patient based on the content of the related information and the relationships between the nodes defined in each path. The process of identifying a path will be described later.

[0049] Note that the information in the node may be updated depending on the content of the medical data, the path identification, etc., and a node that prompts the user for input may be added. As an example, suppose that "blood pressure drop" is recorded as the patient's blood pressure value included in the "vital signs" item collected as related information r12 in node n12 in Figure 2.

[0050] If more detailed information such as "systolic blood pressure drop of 30 mmHg or more" is obtained from measuring the patient's blood pressure, the additional information "systolic blood pressure drop of 30 mmHg or more" can be recorded in addition to the "blood pressure drop" entry recorded in node n12.

[0051] Alternatively, a node for inputting blood pressure details such as "systolic blood pressure drop" may be provided near node n12, and when that information is obtained from the patient through additional testing, the blood pressure details may be recorded in that node. Also, the destination of the connected node may be switched in response to the changing condition of the patient or newly obtained medical data.

[0052] For example, if a patient's blood pressure measurement reveals that the blood pressure, which initially dropped, has stabilized at a normal value, the entry for "dropped blood pressure" in the "vital" section can be changed to "normal blood pressure." In this case, the connections between nodes can also be changed from a node connection relationship suitable for a patient with "dropped blood pressure" to a node connection relationship suitable for a patient with "normal blood pressure."

[0053] Furthermore, clinical card CC2 has a node n21 for "dyspnea", a node n22 for "SPO2 decrease", a node n23 for "oxygen concentration change", a node n24 for "stable breathing", and a node n25 for "stable SPO2".

[0054] Furthermore, in the clinical card CC2 of FIG. 2, the chronological order relationship between nodes is defined, such as "node n21 → node n23 → node n24" or "node n22 → node n23 → node n25."

[0055] In addition, clinical card CC2 has set thereon related information r12 representing "vital signs" related to node n12, related information r13 representing "dialysis setting values" related to node n13, related information r14 representing "observation items" related to node n14, and related information r15 representing "vital signs" related to node n15.

[0056] 1, the description will be continued. The collection function 127c collects the related information set in the clinical card. For example, the collection function 127c collects the related information set in each node of the clinical card identified by the card identification function 127b.

[0057] As an example, when the card identification function 127b identifies the clinical card CC1 in Figure 2, the collection function 127c collects information on "vital signs," "dialysis setting values," and "observation items" set as related information r12 to r15.

[0058] Furthermore, the related information collected by the collection function 127c includes the time representing the related information (for example, if the related information is "vital signs," the time the vital signs were measured). This allows the collection function 127c to collect appropriate related information based on the time representing the related information and the chronological order relationship between the nodes defined in the clinical card, even when the content is the same "vital signs" as in related information r12 and related information r15.

[0059] When the clinical card is identified, if there is related information that cannot be collected due to reasons such as non-measurement, the collection function 127c may monitor the patient information DB 111 of the information integration system 110 via the NW interface 121 and the network 130. In this case, when information corresponding to the related information is added to the patient information DB 111, the collection function 127c collects the information as related information.

[0060] The collection function 127c may also collect related information based on an image of the medical professional captured by the camera 125 or the voice of the medical professional captured by the microphone 126. The collection function 127c may also monitor a storage device provided in an inspection device (not shown) or the like connected via the network 130. In this case, when information corresponding to related information is stored in the storage device of the inspection device, the collection function 127c collects the information as related information.

[0061] The path identification function 127d identifies paths that connect nodes and define the relationships between the nodes based on multiple clinical cards and the time series of performed clinical tasks. This process connects multiple clinical cards in a manner appropriate for the condition of each individual patient, allowing the construction of dedicated network nodes that are useful for assessing the condition of each individual patient and predicting medical procedures.

[0062] For example, the path identification function 127d identifies the condition of the target patient from the related information collected by the collection function 127c. The node corresponding to the identified patient condition changes from an unidentified node to a specific node. Also, for example, the path identification function 127d identifies a clinical task performed on the target patient from the related information collected by the collection function 127c. The node corresponding to the identified clinical task changes from an unidentified node to a specific node.

[0063] Furthermore, for example, the path identification function 127d identifies a path representing the context of the target patient based on multiple nodes corresponding to the identified patient's condition or clinical task, and the causal relationships between the nodes, the chronological order relationships between the nodes, the clinical relationships such as IF-then between the nodes, etc., specified in each path of the clinical card.

[0064] In addition to the degree of association such as causality, paths corresponding to the patient's condition, diagnosis, medical action, and planned medical action may be specified based on the SOAP syntax. The SOAP syntax is a recording method for organizing the patient's condition and medical action according to "Subject (subjective information)," "Object (objective information)," "Assessment," and "Plan."

[0065] In this case, for example, a node relating to a patient's vital signs represents objective information, and a node relating to a diagnosis or assessment of the patient's condition inferred or determined from the vital signs represents an evaluation, and both nodes can be connected as a path linking objective information and evaluation.

[0066] Whether the path is a causal relationship, a chronological relationship, or the like, the paths between nodes are defined in advance based on the type of item represented by the node and are designed within the clinical card. Of the multiple nodes corresponding to the condition or clinical task of the target patient, a node to which information has been input is identified as a path representing the context of the target patient, and changes from an unidentified path to a specified path.

[0067] The path identification function 127d may identify paths in chronological order or in reverse order. For example, when a node representing a past state of the target patient and a node representing a current state of the target patient are identified, the path identification function 127d may identify a path leading to the node representing the current state of the target patient in chronological order by identifying the state of the target patient next to the node representing the past state of the target patient or the medical procedure performed on the target patient next to the node representing the past state of the target patient.

[0068] Also, for example, when a node representing the current condition of the target patient is identified, the path identification function 127d may identify the path in reverse order by identifying the condition of the target patient before the node or the medical procedure performed on the target patient before the node.

[0069] 3 is a diagram illustrating an example of path identification processing. As a premise, it is assumed that the clinical card CC1 of “dialysis” and the clinical card CC2 of “artificial ventilator” shown in FIG. 2 have been identified by the card identification function 127b.

[0070] It is also assumed that the collection function 127c has collected related information r12 to r15 and r22 to r25.

[0071] 3, the path identification function 127d identifies that the blood pressure of the target patient has decreased based on the blood pressure measurement value (hereinafter also referred to as blood pressure value) of the target patient included in the "vital" of the collected related information r12. As a result, the path identification function 127d can identify that the condition of the target patient is "decreased blood pressure" represented by node n12.

[0072] Furthermore, the path identification function 127d identifies that the medical staff has stopped the water removal that they had been performing during dialysis from the information included in the "dialysis setting value" of the collected related information r13. This allows the path identification function 127d to identify that "water removal stop" represented by node n13 has been performed on the target patient.

[0073] Generally, if a patient undergoing dialysis experiences a drop in blood pressure, the water removal is stopped. For this reason, in the example of Figure 3, the path connecting node n12 and node n13 defines the causal relationship between "drop in blood pressure" and "stop water removal."

[0074] In the example of Figure 3, the path identification function 127d identifies from the collected related information that node n12 represents "blood pressure drop" and node n13 represents "water removal discontinued." From this, the path identification function 127d determines that a causal relationship defined in the path connecting node n12 and node n13 is recognized. Therefore, the path identification function 127d identifies the path connecting node n12 and node n13 as a path representing the context of the target patient.

[0075] The path connecting node n12 and node n13 may have a chronological order relationship defined such that "water removal is stopped" after "blood pressure drop" is recognized. In this example, when the path identification function 127d determines from the collected related information that a chronological order relationship is recognized between node n12 and node n13, it identifies the path as a path representing the context of the target patient.

[0076] Furthermore, a predetermined IF-then rule such as "if 'blood pressure drop' is detected, 'stop water removal'" may be defined in the path connecting node n12 and node n13.

[0077] In this example, when the path identification function 127d determines from the collected related information that a relationship expressed by an IF-then rule exists between node n12 and node n13, it identifies the path connecting node n12 and node n13 as a path representing the context of the target patient. In this case, the path identification between nodes functions as a connection from a node representing an evaluation to a node representing a plan.

[0078] The path identification function 127d may also connect paths between nodes by identifying that the target patient is experiencing symptoms of respiratory distress from the contents of the "observation items" in the collected related information r14. In this case, the path identification between nodes functions as a connection from a node representing objective information to a node representing evaluation. This allows the path identification function 127d to identify that the target patient's condition is "respiratory distress" represented by node n14.

[0079] Generally, when fluid accumulates in the lungs or chest, it can cause shortness of breath and difficulty breathing, so stopping the fluid removal performed during dialysis can cause the patient to experience symptoms of respiratory distress. For this reason, the path connecting node n13 and node n14 defines the causal relationship between "stopping fluid removal" and "dyspnea."

[0080] In the example of Fig. 3, the path identification function 127d identifies from the collected related information that node n13 represents "water removal discontinued" and node n14 represents "dyspnea." From this, the path identification function 127d determines that a causal relationship defined in the path connecting node n13 and node n14 is recognized. Therefore, the path identification function 127d identifies the path connecting node n13 and node n14 as a path representing the context of the target patient.

[0081] Furthermore, the path identification function 127d identifies that node n14 of clinical card CC1 and node n21 of clinical card CC2 both represent "dyspnea," and that "dyspnea" is a node common to clinical card CC1 and clinical card CC2 (hereinafter also referred to as a common node). From this, the path identification function 127d identifies that the path connecting node n13 and node n14 is also the path connecting node n13 and node n21.

[0082] For example, even if "dyspnea" cannot be identified from the related information r21 set in clinical card CC2, the path identification function 127d may identify that node n21 represents "dyspnea" because it has been identified that node n14 represents "dyspnea" and because clinical card CC2 for "artificial ventilator" has been identified.

[0083] By linking peripheral information in this way and connecting it to a node that evaluates the patient's condition, it is possible to inform the user of an evaluation of the patient's condition that the user himself or herself may not be aware of.

[0084] Furthermore, the path identification function 127d identifies that the oxygen concentration of the ventilator has been changed from the oxygen concentration setting value included in the "ventilator setting value" of the collected related information r23. This enables the path identification function 127d to identify that the clinical task of "changing oxygen concentration" has been performed by a medical professional.

[0085] Generally, shortness of breath is caused by a lack of oxygen, so in the case of patients wearing a ventilator, measures are taken to increase the oxygen concentration setting of the ventilator. For this reason, the path connecting node n21 and node n22 defines the causal relationship between "shortness of breath" and "change in oxygen concentration."

[0086] In the example of Fig. 3, the path identification function 127d identifies from the collected related information that node n21 represents "dyspnea" and node n22 represents "oxygen concentration change." From this, the path identification function 127d determines that a causal relationship defined in the path connecting node n21 and node n22 is recognized. Therefore, the path identification function 127d identifies the path connecting node n21 and node n22 as a path representing the context of the target patient.

[0087] The path identification function 127d identifies paths connecting nodes as described above, thereby generating context information specific to the target patient, in which nodes representing the condition of the target patient or clinical tasks performed on the target patient are connected by paths. The generated context information is stored in the context DB 122b in association with the patient ID of the target patient.

[0088] The path identification function 127d may identify a path representing the context of the target patient by taking into account the temporal relationship. For example, even if a correlation is recognized between nodes, the path identification function 127d can determine whether to identify a path by taking into account the temporal relationship between the nodes. For example, if the temporal relationship between events represented by two nodes is distant, the path connecting the nodes may not be identified as a path representing the context of the target patient.

[0089] This is because even if a causal relationship or the like is recognized between the nodes of the events themselves, if the time interval between the nodes is quite far and the temporal relationship is distant, the connection between the two is considered to be not very important. If even such things that are considered to be of low importance are identified as paths representing the context of the target patient, the number of paths may increase, making it difficult for the user to visually grasp the context of the target patient.

[0090] For example, when defining paths between clinical cards, the degree of relationship between nodes may be defined numerically, and the degree of relationship may be multiplied by a coefficient corresponding to the time interval between the nodes, so that the longer the time interval, the lower the relationship value multiplied by the coefficient. In this case, paths with relationship values ​​above a predetermined threshold are identified as valid paths.

[0091] In addition, a path may not only connect nodes within the same clinical card, but also connect a node within one clinical card to a node within another clinical card. In this case, the connected clinical cards may be considered as a single larger node that includes multiple nodes.

[0092] Here, Fig. 4 is a diagram illustrating another example of the path identification process. In Fig. 4, it is assumed that a clinical card CC3 for "antipyretic," a clinical card CC4 for "sleeping pill," a clinical card CC5 for "delirium," and a clinical card CC6 for "sepsis" are additionally identified by the card identification function 127b in addition to the example in Fig. 3.

[0093] Clinical card CC3 has a node n31 for "antipyretic medication" and a node n32 for "no fever." Clinical card CC4 has a node n41 for "hypnotic medication" and a node n42 for "sleep state evaluation." Clinical card CC5 has a node n51 for "delirium evaluation" and a node n52 for "wakefulness monitoring."

[0094] Furthermore, clinical card CC6 has a node n61 for "decreased fluid removal", a node n62 for "infectious disease B", a node n63 for "fever", a node n64 for "infectious disease A", and a node n65 for "no fever".

[0095] 4, the path identification function 127d determines that a relationship defined in the path is recognized between node n42 ("Sleep State Assessment") of clinical card CC4 of "Sleeping Pills" and node n51 ("Delirium Assessment") of clinical card CC5 of "Delirium." Also, it is assumed that a predetermined period of time has elapsed between the time when the "Sleep State Assessment" was performed on the target patient and the time when the "Delirium Assessment" was performed on the target patient.

[0096] In addition, in the example of Figure 4, even if a relationship defined in a path is recognized between nodes, if the temporal relationship between the nodes is distant, the path identification function 127d will not identify the path connecting the nodes as a path representing the context of the target patient.

[0097] 4, as described above, a predetermined period of time or more has elapsed between the time the "sleep state assessment" was performed on the target patient and the time the "delirium assessment" was performed on the target patient. Therefore, the path identification function 127d does not identify the path connecting node n42 and node n51 (the path indicated by the dashed line in FIG. 4) as a path representing the context of the target patient.

[0098] 4, node n11a represents "edema," and a path is defined between node n11a and clinical card CC6 of "sepsis." In addition, in the example of Fig. 4, the path identification function 127d determines that a relationship defined in the path is recognized between node n11a and clinical card CC6, and identifies the path connecting node n11a and clinical card CC6 as a path representing the context of the target patient.

[0099] In this case, the user can understand that the target patient may have developed "sepsis" due to "edema."

[0100] In the example of Figure 4, node n61 "Decreased water removal" indicates that the amount of water removed from the target patient has decreased, and a path is defined between node n61 and node n62 of "Infectious disease B." "Infectious disease B" indicates that the target patient is infected with infectious disease B.

[0101] Also, in the example of Figure 4, the path identification function 127d determines that a relationship defined in the path is recognized between node n61 and node n62, and identifies the path connecting node n61 and node n62 as a path representing the context of the target patient.

[0102] In this case, the user can understand from the chronological order of the nodes "Decreased Water Removal" → "Infectious Disease B" that the patient may have contracted Infectious Disease B due to a decrease in the amount of water removed.

[0103] 4, the "fever" of node n63 indicates that the target patient has a fever, and a path is defined between node n62 and node n63. In the example of Fig. 4, the path identification function 127d determines that a relationship defined in the path is recognized between node n62 and node n63, and identifies the path connecting node n62 and node n63 as a path representing the context of the target patient.

[0104] In this case, the user can understand that the target patient developed a fever as a result of contracting infectious disease B from the chronological order of the nodes "infectious disease B" → "fever."

[0105] 4, the "antipyretic medication" of node n31 indicates that an antipyretic medication has been administered to the target patient, and a path is defined between node n63 and node n31. In the example of Fig. 4, the path identification function 127d determines that a relationship defined in the path is recognized between node n63 and node n31, and identifies the path connecting node n63 and node n31 as a path representing the context of the target patient.

[0106] In this case, the user can understand from the chronological order relationship of the nodes "Fever" → "Administration of Antipyretic" that the patient was administered an antipyretic as a treatment for the fever caused by infectious disease B.

[0107] 4, "no fever" in node n32 indicates that the target patient does not have a fever, and a path is defined between node n31 and node n32. In the example of Fig. 4, the path identification function 127d determines that a relationship defined in the path is recognized between "antipyretic medication" and "no fever," and identifies the path connecting node n31 and node n32 as a path representing the context of the target patient.

[0108] In this case, the user can understand from the chronological relationship of the nodes "administered antipyretic" → "no fever" that the fever subsided because the patient was administered an antipyretic. Furthermore, from the series of events, the user can understand the context of the patient, that is, the patient had a fever due to infection B, but the fever subsided after being administered an antipyretic.

[0109] In addition, in the example of Fig. 4, a path is defined between the clinical card CC2 of "artificial ventilator" and the node n52 of "awake monitoring." In the example of Fig. 4, the path identification function 127d does not identify this path as a path representing the context of the target patient.

[0110] However, for example, if the collection function 127c collects related information that can derive the relationship between clinical card CC2 and node n52, the path identification function 127d will identify the path as a path representing the context of the target patient.

[0111] The path identification function 127d may identify a path representing the context of the target patient using a trained model that has learned the connection relationships between nodes using a known machine learning technique (including deep learning technique). For example, the trained model is a trained model that infers, from information representing a first node, information representing a second node that is in a connection relationship with the first node.

[0112] As an example, the trained model is a trained model that learns the relationship between the two using a training dataset in which information representing a first node and related information about the first node are used as input training data, and information representing a second node is used as output training data.

[0113] The path identification function 127d inputs information representing the first node and related information of the first node into the trained model, and identifies a path representing the context of the target patient based on the output result of the trained model.

[0114] 1, the explanation will be continued. The presentation function 127e presents context information that indicates the relationship between nodes connected by paths. For example, the presentation function 127e presents the context information to the user by displaying the nodes represented by solid lines in FIGS. 3 and 4 and the paths connecting the nodes represented by thick lines on the display 124 in a manner that allows the chronological order relationship between the nodes to be understood.

[0115] Furthermore, for example, when a common node exists, the presentation function 127e, when presenting the context information, overlaps the common nodes and displays them as a single node on the display 124. This allows the context information to be presented to the user as a single graph even when multiple clinical cards are involved in the context information.

[0116] The presentation function 127e may extract and present a portion of the context information in accordance with a user instruction. In this case, the presentation function 127e receives a selection input of one node from the nodes presented as context information from the user. Then, the presentation function 127e extracts and presents nodes and paths located within a predetermined range of the selected node (for example, within a range of two nodes before and after the selected node).

[0117] By extracting and presenting a portion of the context information in accordance with the user's instructions in this way, the user can more easily understand the patient condition or medical procedure that is closely related to the patient condition or medical procedure that the user wishes to focus on.

[0118] The presentation function 127e may also automatically extract nodes of high importance or nodes around clinical cards.

[0119] In this case, the presentation function 127e determines the level of importance based on indicators such as the number of related information collected as related information for the node, the number of medical professionals who have recorded information about the node in electronic medical records, etc., the number of times information about the node has been viewed, whether the node or clinical card is related to the main illness of the target patient, and the number of nodes in the clinical card that represent highly urgent treatments.

[0120] By extracting and presenting the important parts of the context information in this way, the user can easily understand the patient conditions and medical procedures that are highly relevant to the patient conditions and medical procedures that are of high importance.

[0121] In addition, when extracting and presenting a portion of the context information, the presentation function 127e may display a reduced image of the entire context information together with information indicating which portion of the context information has been extracted.

[0122] This makes it easier for the user to understand which part of the entire context information the extracted part corresponds to.

[0123] Furthermore, the presentation function 127e presents related information together with the context information. For example, the presentation function 127e receives a selection input of one node from the nodes presented as context information from the user. Then, the presentation function 127e displays the context information on the left side of the display 124 and the related information defined for the selected node on the right side.

[0124] The presentation function 127e may display not only the related information of the selected node but also the related information of nodes located within a predetermined range of the selected node. In this case, the presentation function 127e may display the related information in accordance with the chronological order relationship between the nodes.

[0125] By presenting related information together as described above, the user can easily understand the basis on which the nodes are connected to each other.

[0126] The operation of each function will be described below with reference to Figures 5 to 7. Figure 5 is a diagram illustrating an example of the operation of each function of the medical information processing device 120. In the example of Figure 5, it is assumed that the attending physician HW1, ICU doctor HW2, ICU nurse HW3, and ICU doctor HW4 are in charge of caring for the target patient.

[0127] For example, when it is decided that a target patient will be admitted to the ICU, the attending physician HW1 inputs information representing the patient's current condition to the medical information processing device 120 ((1)-1). Specifically, the attending physician HW1 uses the input interface 123 to input information representing "respiratory failure" and "renal failure" as the patient's current condition.

[0128] Next, the acquisition function 127a accepts input of information representing "respiratory failure" and "renal failure" via the input interface 123. The card identification function 127b refers to the clinical card DB 122a and identifies a clinical card CC7 corresponding to "respiratory failure" and a clinical card CC8 corresponding to "renal failure."

[0129] The collection function 127c then automatically collects the relevant information ((1)-2).

[0130] Specifically, the collection function 127c collects "information on respiratory condition (dyspnea)" and "information on transport condition (emergency transport)" defined as related information r7 on the clinical card CC7 from the patient information DB 111 of the information integration system 110 via the NW interface 121 and the network 130. Similarly, the collection function 127c collects "information on dialysis frequency (dialysis twice a week)" defined as related information r8 on the clinical card CC7.

[0131] Next, the ICU doctor HW2 in charge of the day shift inputs a clinical task to the medical information processing device 120 ((2)-1). Specifically, the ICU doctor HW2 inputs information representing "artificial respiration" as a clinical task corresponding to "respiratory failure." Similarly, the ICU doctor HW2 inputs information representing "artificial dialysis" as a clinical task corresponding to "renal failure."

[0132] Next, the card identification function 127b refers to the clinical card DB 122a and identifies the clinical card CC2 corresponding to "artificial respiration" and the clinical card CC1 corresponding to "artificial dialysis."

[0133] Next, the path identification function 127d automatically identifies a path ((2)-2).

[0134] Specifically, the path identification function 127d determines that a path-defined relationship exists between clinical card CC7 and clinical card CC2 based on the related information r7 (dyspnea, emergency transport) of the clinical card collected by the collection function 127c, and the information "respiratory failure" and "artificial ventilator." Therefore, the path identification function 127d identifies the path between clinical card CC7 and clinical card CC2 as a path representing the context of the target patient.

[0135] Similarly, the path identification function 127d determines that a path-defined relationship is recognized between clinical card CC7 and clinical card CC2 based on the clinical card related information r8 (dialysis twice a week) collected by the collection function 127c, and the information "renal failure" and "dialysis." Therefore, the path identification function 127d identifies the path between clinical card CC7 and clinical card CC2 as a path representing the context of the target patient.

[0136] The collection function 127c then automatically collects the relevant information ((2)-3).

[0137] Specifically, the collection function 127c collects collectable information from among "blood gas values," "ventilator settings," and "observation items" defined as related information r22 to r25 on the clinical card CC2. Similarly, the collection function 127c collects collectable information from among "vital values," "dialysis settings," and "observation items" defined as related information r12 to r15 on the clinical card CC1.

[0138] Next, the path identification function 127d automatically identifies the clinical task performed on the target patient ((2)-4). Specifically, the path identification function 127d identifies that water removal is being performed during dialysis based on the related information r13 (dialysis setting values) of the clinical card collected by the collection function 127c. In other words, the path identification function 127d identifies the node n13a representing "water removal" as the clinical task performed on the target patient.

[0139] Next, ICU nurse HW3, who is in charge of the day shift, performs clinical tasks ((3)-1). Specifically, ICU nurse HW3 measures vital signs, blood gases, and observes the patient's respiratory condition.

[0140] The collection function 127c automatically collects related information ((3)-2). Specifically, the collection function 127c monitors the patient information DB 111, and when information corresponding to the related information defined in the clinical card CC1 and the clinical card CC2 is added, the collection function 127c collects the information as related information. In the example of FIG. 5, the collection function 127c collects information corresponding to "vital signs," "blood gas values," and "observation items" as related information.

[0141] Next, the path identification function 127d identifies the patient's condition ((4)-1). Specifically, the path identification function 127d identifies that the target patient's blood pressure has decreased from the blood pressure measurement results included in the collected "vitals." In other words, the path identification function 127d identifies node n12, which represents "decreased blood pressure," as the patient's condition.

[0142] Next, the ICU doctor HW4, who is in charge of the night shift, performs a clinical task ((4)-2). Specifically, the ICU doctor HW4 stops the water removal during dialysis. After this, the collection function 127c collects information that the water removal during dialysis has been stopped as information included in the "dialysis setting value" of the related information r13.

[0143] Next, the path identification function 127d identifies a clinical task ((4)-3). Specifically, the path identification function 127d identifies “water removal stop” as the clinical task performed on the target patient based on the collected information that water removal during dialysis has been stopped.

[0144] Next, the ICU doctor HW4 performs clinical tasks ((4)-4). Specifically, the ICU doctor HW4 observes the patient's respiratory condition and changes the oxygen concentration setting of the ventilator.

[0145] Thereafter, the collection function 127c collects information that the target patient is experiencing respiratory distress as information included in the "Observation Items" of the related information r14 and r24. The collection function 127c also collects information that the oxygen concentration setting has been changed, which is included in the "Ventilator Setting Value" of the related information r13.

[0146] Next, the path identification function 127d identifies the patient's condition and clinical task ((4)-5). Specifically, the path identification function 127d identifies "dyspnea" as the condition of the target patient from the collected information that the target patient is in a state of dyspnea. Also, from the collected information that the oxygen concentration setting of the ventilator has been changed, the path identification function 127d identifies "oxygen concentration change" as the clinical task performed on the target patient.

[0147] Next, the path identification function 127d identifies a path ((4)-6). Specifically, the path identification function 127d determines that a relationship defined in the path is recognized between node n12 and node n13 based on "blood pressure decrease" and "water removal discontinued." Therefore, the path identification function 127d identifies the path between node n12 and node n13 as a path representing the context of the target patient.

[0148] Similarly, the path identification function 127d determines that a relationship defined in the path is recognized between node n13 and node n14 based on "water removal discontinued" and "dyspnea." Therefore, the path identification function 127d identifies the path between node n13 and node n14 as a path representing the context of the target patient.

[0149] Furthermore, the path identification function 127d identifies that the nodes n14 and n21 are a common node representing “dyspnea.” The path identification function 127d identifies that the path connecting the nodes n13 and n14 is also the path connecting the nodes n13 and n21.

[0150] Furthermore, the path identification function 127d determines that a relationship defined in the path is recognized between node n21 and node n23 based on "dyspnea" and "oxygen concentration change." Therefore, the path identification function 127d identifies the path between node n21 and node n23 as a path representing the context of the target patient.

[0151] In this way, each function of the medical information processing device 120 generates context information for the target patient that represents the chronological order of each node: "drop in blood pressure" → "cessation of fluid removal" → "dyspnea" → "change in oxygen concentration."

[0152] Next, the use of context information will be described. Figures 6 and 7 are diagrams for explaining an example of the use of context information. In Figure 6, it is assumed that on the second day since the target patient in the example of Figure 5 was admitted to the ICU, the ICU doctor HW2 on day shift is considering weaning the target patient from the ventilator.

[0153] If context information had not been generated, ICU physician HW2 would have had to perform tasks such as checking the current ventilator settings and investigating why the oxygen concentration setting was changed by a medical professional other than himself.

[0154] When multiple medical professionals are responsible for caring for a single patient, the information related to medical procedures is recorded by each medical professional, and the location where the information is recorded may differ depending on the medical professional. In this case, ICU physician HW2 will have to search through various recording locations, which is likely to take a long time to collect the necessary information.

[0155] Furthermore, even if information were recorded in the same place, the volume of information recorded would increase because many medical professionals are involved in care, and it could take a long time to find the information you need.For this reason, if context information had not been generated, ICU physician HW2 would have had to perform tedious work to understand who performed what medical procedure and why.

[0156] In contrast, in this embodiment, the ICU doctor HW2 can easily understand that "the oxygen concentration setting value was changed by a medical professional other than himself because the target patient became short of breath" by focusing on the AR1 portion of the context information shown in Figure 6.

[0157] Also, in Figure 7, it is assumed that ICU doctor HW2 wants to investigate the cause of the patient's worsening respiratory condition. In this case, by focusing on part AR2 of the context information shown in Figure 7, ICU doctor HW2 can easily understand that "water removal was originally performed during dialysis, but a drop in blood pressure was observed, so water removal was stopped" and "the respiratory condition worsened as a result of water removal being stopped."

[0158] Furthermore, by understanding the above, the ICU doctor HW2 can decide on a policy of "first removing fluid and solving the problem of hypotension." In this way, the medical information processing device 120 according to this embodiment can support efficient decision-making on a care policy by presenting the user with context information specific to the target patient.

[0159] Next, a description will be given of the processing executed by the medical information processing apparatus 120. Fig. 8 is a flowchart showing an example of the processing executed by the medical information processing apparatus 120.

[0160] First, the acquisition function 127a receives input of information representing the condition of a patient (step S101). For example, the acquisition function 127a receives input of information representing the current condition of a patient from a user via the input interface 123.

[0161] Next, the card identification function 127b identifies a clinical card corresponding to the patient's condition (step S102). For example, the card identification function 127b refers to the clinical card DB 122a in the memory circuit 122 and identifies a clinical card corresponding to the patient's current condition accepted in step S101.

[0162] Next, the acquisition function 127a receives an input of a clinical task (step S103). For example, the acquisition function 127a receives an input of information representing a clinical task for the current condition of the patient received in step S101 from the user via the input interface 123.

[0163] Next, the card identification function 127b identifies a clinical card corresponding to the clinical task (step S104). For example, the card identification function 127b refers to the clinical card DB 122a in the storage circuit 122 and identifies a clinical card corresponding to the information representing the clinical task received in step S103.

[0164] Next, the collection function 127c collects related information (step S105). For example, the collection function 127c refers to the patient information DB 111 of the information integration system 110 and collects related information defined in the clinical card corresponding to the current condition of the patient identified in step S102. Similarly, the collection function 127c collects related information defined in the clinical card corresponding to the clinical task identified in step S104.

[0165] It should be noted that the process of collecting relevant information defined in the clinical card corresponding to the patient's current condition may be performed after step S102 and before step S103.

[0166] Next, the path identification function 127d identifies the patient's condition represented by the node on the clinical card based on the related information (step S106). For example, the path identification function 127d derives the patient's condition from the related information collected in step S105 and identifies the patient's condition represented by the node on the clinical card identified in steps S102 and S104.

[0167] Next, the path identification function 127d identifies clinical tasks represented by nodes on the clinical cards based on the related information (step S107). For example, the path identification function 127d derives clinical tasks from the related information collected in step S105 and identifies clinical tasks represented by nodes on the clinical cards identified in steps S102 and S104.

[0168] The process of step S107 may be performed in parallel with the process of step S106, or may be performed after step S105 and before step S106.

[0169] Next, the path identification function 127d identifies paths between nodes (step S108). For example, the path identification function 127d identifies paths between nodes to be adopted as paths representing the context of the target patient based on the related information collected in step S105 and the relationships between nodes defined in the paths. This generates context information of the target patient.

[0170] Next, the presentation function 127e presents the generated context information of the target patient (step S109). For example, the presentation function 127e presents the context information of the target patient to the user by displaying, on the display 124, nodes and paths within a predetermined range arranged around a node selected by the user, together with related information of the node.

[0171] Next, the processing circuitry 127 stores the generated context information of the target patient in the context DB 122b of the storage circuitry 122 (step S110), and ends this process. For example, the processing circuitry 127 associates the generated context information of the target patient with the patient ID of the target patient and stores them in the context DB 122b. Note that the processing of step S110 may be performed after step S108 and before step S109.

[0172] As described above, the medical information processing device 120 of this embodiment identifies paths connecting nodes having relationships defined on each of the clinical cards based on multiple clinical cards in which nodes define the relationship between the condition of the target patient, clinical tasks representing medical procedures performed for that condition, and the patient's condition changing as a result of the performance of the clinical tasks, and the timeline of the clinical tasks performed on the patient, and presents context information representing the relationship between the nodes connected by the paths.

[0173] This allows the user to understand, for example, the condition of the target patient, the treatment performed for that condition, and how the treatment changed the target patient's condition in chronological order. Therefore, even if a medical professional other than the user performed treatment on the target patient, the user can easily understand why the treatment was performed and how the target patient's condition changed as a result, without having to collect various data about the target patient that is stored in a scattered manner or organize the collected data. In other words, the medical information processing device 120 according to this embodiment can efficiently support the determination of the treatment context for each subject.

[0174] The above-described embodiment can be modified as needed by partially changing the configuration or functions of each device. Therefore, several modifications of the above-described embodiment will be described below as other embodiments. The following mainly focuses on differences from the above-described embodiment, and detailed descriptions of commonalities with the content already described will be omitted. The modifications described below may be implemented individually or in appropriate combination.

[0175] (Variation 1) In the above embodiment, a form was described in which a path representing the context of a target patient is identified from the related information defined in a clinical card and the relationship between nodes. In this modified example, a form will be described in which a path representing the context of a target patient is identified by further taking into account the attributes defined in the nodes.

[0176] In this modification, attributes are defined for the clinical card and the nodes arranged on the clinical card. The attributes are defined, for example, by the classification of the SOAP method. Generally, the classification of the SOAP method is as follows:

[0177] S (subjective): Subjective information (information obtained from what the patient said, etc.) O (objective): Objective information (objective information obtained from examinations, tests, etc.) A (assessment): Evaluation (comprehensive evaluation based on the doctor's diagnosis and analysis and interpretation of the contents of O and S) P (plan): Plan (treatment) (treatment policy and content decided based on A, lifestyle guidance, etc.)

[0178] As an example, attributes are defined as follows: "dyspnea" is "S," "decreased SPO2" is "O," "hypoxemia" is "A," and "change in oxygen concentration" is "P." In this modification, the path identification function 127d identifies a path that represents the context of the target patient so that the node attributes are linked in the form of "S" → "O" → "A" → "P."

[0179] For example, when the path identification function 127d identifies the nodes "S" and "O," it identifies the path representing the context of the target patient by inferring the nodes "A" → "P" that follow from "S" → "O" from the related information and causal relationships specified in the clinical card.

[0180] Also, for example, if the path identification function 127d is unable to identify the "A" node, it identifies the path representing the context of the target patient by inferring the "A" node leading to the "P" node from the related information and causal relationships specified in the clinical card.

[0181] Furthermore, the path identification function 127d may use a known machine learning technique to identify a path representing the context of the target patient using a trained model that has learned the relationships between "S" → "O" and "A" → "P." For example, the trained model is a trained model that infers "A" → "P" from "S" → "O."

[0182] As an example, the trained model is a trained model that learns the relationship between the two using a training dataset in which "S" → "O" and the related information of each node representing "S" → "O" are used as input training data, and "A" → "P" is used as output training data.

[0183] The path identification function 127d inputs "S" → "O" and the related information of each node representing "S" → "O" into the trained model, and identifies the path representing the context of the target patient based on "A" → "P" output by the trained model.

[0184] According to this modification, even if the SOAP is not organized and recorded, it is possible to present to the user the context information of the target patient in which the SOAP is organized based on the connection relationships between the nodes. Also, even if the SOAP is organized and recorded, the user may be able to grasp missing elements (e.g., A, which was not considered when P was planned) from the connection relationships between the nodes.

[0185] (Variation 2) In the above-described embodiment, a form has been described in which context information representing the relationship between nodes representing the past states of a target patient or clinical tasks performed in the past up to a node representing the present (a node representing the current state of the target patient or a node representing the clinical task most recently performed) is presented. In this modified example, a form will be described in which nodes representing the future state of a target patient or clinical tasks to be performed in the future are predicted and presented.

[0186] In this modification, the path identification function 127d predicts, from a node identified as the state of a patient or a clinical task performed on a patient, the next node that is connected to that node by a path.

[0187] For example, the path identification function 127d identifies the node with the highest probability of being connected to the next path based on statistical values, using context information generated using the same type of nodes or clinical cards as the clinical cards or nodes used to generate the context information of the target patient as similar cases, and predicts the node as the node that can be connected to the node with the next path.

[0188] Furthermore, the path identification function 127d may predict subsequent nodes that will be connected by a path based on the predicted node.

[0189] The path identification function 127d may predict the next node connected by a path using a trained model that has learned the relationship between a certain node and the node connected next to the certain node by a path using a known machine learning technique. For example, the trained model is a trained model that infers information representing a certain node from information representing the certain node to information representing the node connected next to the certain node by a path.

[0190] As an example, the trained model is a trained model that learns the relationship between the two using a training dataset in which information representing a node and related information about the node are used as input training data, and information representing the node and the next node connected by a path are used as output training data.

[0191] The path identification function 127d inputs information representing the current node and related information of the current node into the trained model, and predicts the next node to be connected by a path based on the output result of the trained model.

[0192] In addition, in this modified example, if there is no node that can be connected next by a path among the nodes of the currently identified clinical card, the card identification function 127b identifies a clinical card that has a node that can be connected next by a path using the same method as described above.

[0193] In this modification, the presentation function 127e presents the generated context information and the prediction result of the next node connected by the path. For example, the presentation function 127e displays the prediction result on the display 124 in a manner that allows the two to be distinguished from each other, for example, by displaying the prediction result in a color different from that of the context information representing the path up to the node representing the present.

[0194] Furthermore, for example, the presentation function 127e may present information representing related information of the node representing the current state and related information defined for the predicted node in addition to the above, which allows the user to easily understand on what basis the next node connected by the path is predicted.

[0195] According to this modified example, for example, the user can easily understand not only the course of events that the target patient has taken to reach his or her current condition, but also the course of events that the target patient is predicted to take in the future.

[0196] (Variation 3) In the above-described embodiment, a configuration in which context information is generated by combining clinical cards pre-stored in the clinical card DB 122a has been described. In this modified example, a configuration in which a clinical card added by the user or a clinical card with changed content can be used will be described.

[0197] In this modified example, the concept of "inheritance" from the Unified Medical Language System (UMLS) is applied to clinical cards, enabling the efficient addition and modification of clinical cards.

[0198] For example, when registering a new clinical card in the clinical card DB 122a, in this modification, the user selects a card to be a parent card from the clinical cards stored in the clinical card DB 122a. The processing circuit 127 identifies a card similar to the parent card from among the clinical cards stored in the clinical card DB 122a.

[0199] For example, the processing circuitry 127 determines whether each clinical card is similar to a selected parent card based on the similarity of the check flow related to the clinical card, etc. As an example, since the double-check or single-check flow at the time of administration is common regardless of the drug, if a clinical card related to a certain drug is designated as a parent card, the processing circuitry 127 determines that clinical cards related to other drugs are similar to the parent card.

[0200] As another example, if the parent card is a clinical card related to a general decision-making flow as described in the guidelines, the processing circuit 127 determines that other clinical cards related to the general decision-making flow are similar to the parent card.

[0201] As another example, if the parent card is a clinical card related to a standardized flow such as a clinical pathway, the processing circuit 127 determines that other clinical cards related to the standardized flow are similar to the parent card.

[0202] In addition, the processing circuitry 127 may set a clinical card that is determined to be similar to a parent card as a child card of the parent card.

[0203] Next, the processing circuitry 127 defines elements (such as nodes and related information defined in the nodes) common to the parent card and all of the clinical cards identified as clinical cards similar to the parent card as a common part. The processing circuitry 127 sets the defined common part as a common part of a newly added child card that inherits the content of the parent card.

[0204] If the user wants to change an element among the elements defined as the common part, the user may input an instruction to exclude the element from the common part. In this case, the processing circuitry 127 excludes the element from the common part.

[0205] Furthermore, the processing circuitry 127 sets elements that are not defined as common parts of the parent card as unique parts, which are changeable elements of the child card.

[0206] The user creates a child card by changing the content of the set unique part (e.g., the patient's condition that is a condition for performing a clinical task, the content of the clinical task, etc.) to unique content. As an example, in the case of a clinical card related to a clinical pathway, the user changes the content of the observation items, variances, etc. that are defined as related information of the nodes that the clinical card has.

[0207] The processing circuitry 127 stores the created child card in the clinical card DB 122a in association with information indicating the condition of the corresponding patient and the clinical task.

[0208] Since child cards inherit the parent card, if the content of the parent card changes, the content of the child card will also change accordingly. This means that if you want to change the content of the common part between the parent card and the child card, you can change the content of the child card just by changing the content of the parent card.

[0209] According to this modified example, for example, if a medical facility reviews the operation of clinical tasks and it becomes necessary to change the contents of clinical cards or add new clinical cards, the user can perform such tasks efficiently.

[0210] (Variation 4) In the above-described embodiment, a form has been described in which the medical information processing device 120 executes the above-described processing via its own input interface 123, display 124, camera 125, and microphone 126. In this modified example, a form will be described in which the above-described processing is executed by a terminal device used by a medical professional and a medical information processing device communicably connected to the terminal device.

[0211] In this modification, the terminal device has the same configuration as the input interface 123, the display 124, the camera 125, and the microphone 126 described as components of the medical information processing device 120 in the above-described embodiment. For example, the terminal device may be realized by a smartphone, a tablet terminal, a personal computer, a small dedicated device similar to a smartphone, or the like.

[0212] Furthermore, the medical information processing device executes the processes via a terminal device, which processes are executed via the input interface 123, the display 124, the camera 125, and the microphone 126 in the above-described embodiment. For example, the medical information processing device is realized by computer equipment such as a server or a workstation.

[0213] The terminal device and the medical information processing device are connected to each other so that they can communicate with each other via the Internet, for example. In this case, the medical information processing system may be constructed in the form of cloud computing, in which the medical information processing device provides services in response to requests from the terminal device.

[0214] Furthermore, the above-described embodiment has described an example in which the acquisition unit, graph identification unit, node identification unit, path identification unit, and presentation unit in this specification are realized by the acquisition function 127a, card identification function 127b, path identification function 127d, and presentation function 127e of the processing circuit 127, respectively, but the embodiment is not limited to this.

[0215] For example, the acquisition unit, graph identification unit, node identification unit, path identification unit, and presentation unit in this specification may be realized by the acquisition function 127a, card identification function 127b, path identification function 127d, and presentation function 127e of the processing circuit 127, respectively, or the same functions may be realized by hardware only, software only, or a combination of hardware and software.

[0216] In the above-described embodiment, the processing circuitry 127 is implemented by a single processor, but the embodiment is not limited to this. For example, the processing circuitry 127 may be configured by combining multiple independent processors, and each processor may execute a program to implement each processing function. Furthermore, each processing function of the processing circuitry 127 may be implemented by being distributed or integrated as appropriate across a single or multiple processing circuits.

[0217] Furthermore, each processing function of the processing circuitry 127 may be realized by a combination of hardware and software, such as circuits. Although an example in which programs corresponding to each processing function are stored in a single storage circuitry 122 has been described here, the embodiment is not limited to this. For example, the programs corresponding to each processing function may be stored in a distributed manner in multiple storage circuits, and the processing circuitry 127 may read and execute each program from each storage circuit.

[0218] Furthermore, the term "processor" used in the description of the above-mentioned embodiments refers to circuits such as a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), an Application Specific Integrated Circuit (ASIC), a programmable logic device (e.g., a Simple Programmable Logic Device (SPLD), a Complex Programmable Logic Device (CPLD), and a Field Programmable Gate Array (FPGA)).

[0219] Instead of storing the program in a memory circuit, the program may be directly embedded in the processor circuit. In this case, the processor realizes its function by reading and executing the program embedded in the circuit. Furthermore, each processor in this embodiment is not limited to being configured as a single circuit, but may be configured as a single processor by combining multiple independent circuits to realize its function.

[0220] Here, the program executed by the processor is provided by being pre-installed in a ROM (Read Only Memory), a storage circuit, etc. Note that this program may also be provided by being recorded in a computer-readable, non-transitory storage medium such as a CD (Compact Disk)-ROM, a FD (Flexible Disk), a CD-R (Recordable), or a DVD (Digital Versatile Disk) in a format that can be installed or executed by these devices.

[0221] This program may also be provided or distributed by being stored on a computer connected to a network such as the Internet and downloaded via the network. For example, this program may be composed of modules including each of the processing functions described above. In actual hardware, a CPU reads and executes the program from a storage medium such as a ROM, whereby each module is loaded into a main memory and generated on the main memory.

[0222] Furthermore, in the above-described embodiment and modified examples, the components of each device shown in the drawings are functional concepts and do not necessarily have to be physically configured as shown in the drawings. In other words, the specific form of distribution or integration of each device is not limited to that shown in the drawings, and all or part of them can be functionally or physically distributed or integrated in any unit depending on various loads, usage conditions, etc.

[0223] Furthermore, all or any part of the processing functions performed by each device may be realized by a CPU and a program analyzed and executed by the CPU, or may be realized as hardware using wired logic.

[0224] Furthermore, among the processes described in the above-mentioned embodiments and modifications, all or part of the processes described as being performed automatically can be performed manually, or all or part of the processes described as being performed manually can be performed automatically using a known method.In addition, the information including the processing procedures, control procedures, specific names, various data and parameters shown in the above documents and drawings can be changed as desired unless otherwise specified.

[0225] The various data handled in this specification are typically digital data.

[0226] According to at least one of the embodiments described above, it is possible to efficiently assist in determining the treatment context for an individual subject.

[0227] Although several embodiments have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, substitutions, modifications, and combinations of embodiments can be made without departing from the spirit of the invention. These embodiments and their modifications are included within the scope and spirit of the invention, as well as within the scope of the invention and its equivalents as defined in the claims. [Explanation of symbols]

[0228] 100 Medical Information Processing System 120 Medical information processing device 122 Memory circuit 122a Clinical Card DB 127 Processing Circuit 127a Acquisition function 127b Card specific function 127c collection function 127d Path Identification Function 127e Presentation function

Claims

1. a path specifying unit that specifies a path that connects the node that represents the first state and the node that represents the second state and specifies a relationship between the nodes, based on a plurality of graphs that define, with nodes, a relationship between a first state of a subject, a clinical task that represents a medical procedure that is performed for the first state, and a second state of the subject that changes as a result of performing the clinical task, and a time series of the clinical task; a presentation unit that presents context information that indicates a relationship between the nodes connected by the path; A medical information processing device comprising:

2. an acquisition unit that acquires medical data indicating a current condition of the subject or a medical procedure to be performed on the subject; a graph identification unit that identifies the graph having the node corresponding to the condition or the clinical task based on the condition or the medical action indicated in the medical data; Further provided with the path identification unit identifies the node representing the first state among the nodes included in the graph identified by the graph identification unit, and identifies the path connecting the identified node and the node representing the second state in chronological order or in reverse chronological order, with the identified node as a base point; The medical information processing device according to claim 1 .

3. a node specifying unit that specifies a common node indicating a node included in a second graph, which is a graph other than the first graph and which is common to the node included in a first graph among the graphs specified by the graph specifying unit, based on association information for deriving the association, which is defined for each node; the presentation unit presents the context information connecting the first graph and the second graph by overlapping the identified common nodes. The medical information processing device according to claim 2 .

4. the presentation unit presents the related information of the node designated by the user together with the context information. The medical information processing device according to claim 3 .

5. the presentation unit presents, together with the context information, the related information of the node designated by the user and the nodes located within a predetermined range of the node; The medical information processing device according to claim 4 .

6. the presentation unit presents the context information configured from the nodes arranged within a predetermined range starting from the node designated by the user. The medical information processing device according to any one of claims 1 to 5.

7. the first state represents a current state of the subject; the second state represents a state of the subject at a time prior to the present time; the path identification unit identifies the path by identifying the node that represents the second state from the node that represents the first state and the related information of the node; The medical information processing device according to any one of claims 3 to 5.

8. the first state represents a current state of the subject; the second state represents a state of the subject in the future from the present time; the path identification unit identifies the path by predicting the node that represents the second state from the node that represents the first state and the related information of the node; The medical information processing device according to any one of claims 3 to 5.

9. The node has one of the following attributes defined: S (subjective information), O (objective information), A (evaluation), and P (planning). the path identification unit, when the node of any of the attributes is not identified among the plurality of nodes linked in the order of S, O, A, and P, identifies the path of the subject by inferring the node of the attribute that is not identified and is connected to the identified node based on the content of the identified node, the attribute of the identified node, and the related information; The medical information processing device according to any one of claims 3 to 5.

10. a path specifying unit that specifies a path that connects the node that represents the first state and the node that represents the second state and specifies a relationship between the nodes, based on a plurality of graphs that define, with nodes, a relationship between a first state of a subject, a clinical task that represents a medical procedure that is performed for the first state, and a second state of the subject that changes as a result of performing the clinical task, and a time series of the clinical task; a presentation unit that presents context information that represents a relationship between the nodes connected by the path; A medical information processing system comprising:

11. A medical information processing method by a medical information processing device, comprising: a path identifying step of identifying paths that connect the nodes representing the first state and the nodes representing the second state and that define the relationship between the nodes, based on a plurality of graphs in which the relationship between a first state of a subject, a clinical task representing a medical procedure performed for the first state, and a second state of the subject that changes as the clinical task is performed, and a time series of the clinical task; a presentation step of presenting context information representing a relationship between the nodes connected by the path; A medical information processing method including:

Images