JP7784282B2 - Medical support equipment - Google Patents

Description

An embodiment of the present invention relates to a medical support device.

Before and after examining a patient, doctors refer to literature related to medical treatment to (a) update treatment methods, (b) expand treatment options and engage in shared decision-making (SDM) with the patient, and (c) confirm the evidence for and precautions regarding treatment methods.

Here, literature information related to medical treatment includes guidelines, papers, case studies, etc., each of which contains information according to medical specialty or disease. Below, guidelines are described as an example of literature information. Each guideline has different granularity and scope of information depending on the specialty, and using only one guideline may provide insufficient or biased information.

On the other hand, it is time-consuming for doctors to refer to all of the guidelines that may apply to a patient, regardless of their own area of expertise. As a result, doctors may provide treatment by referring to only one guideline that corresponds to their area of expertise, without referring to several potentially applicable guidelines. In this case, there is a risk that doctors will not be able to provide optimal treatment due to a lack of information or bias in the guideline they refer to. Specifically, for example, there is a risk that they will not be able to select the most appropriate treatment method for the patient, or that treatment will not be administered at the optimal time.

Special Publication No. 2011-520195

One of the problems that the embodiments disclosed in this specification and drawings attempt to solve is to present multiple pieces of literature information that may be relevant to a patient while reducing the effort required to refer to literature information related to medical treatment. However, the problems that the embodiments disclosed in this specification and drawings attempt to solve are not limited to the above problem. Issues corresponding to the effects of each configuration shown in the embodiments described below can also be positioned as other problems.

A medical support device according to an embodiment includes a first extraction unit, a second extraction unit, a third extraction unit, and an output unit. The first extraction unit extracts first literature information related to a medical term group included in a patient's medical information from among a plurality of literature information related to medical treatment. The second extraction unit extracts any first term group from among the first term groups corresponding to each itemized description portion in the first literature information based on a first degree of association with the medical term group. The third extraction unit extracts any second term group from among the second term groups corresponding to each itemized description portion in second literature information different from the first literature information based on a second degree of association between the extracted first term group and the medical term group. The output unit outputs the first literature information and the second literature information based on the extracted first term group and the extracted second term group.

FIG. 1 is a block diagram showing a medical assistance device and its peripheral configuration according to the first embodiment. FIG. 2 is a block diagram showing the configuration of the medical assistance device according to the first embodiment. FIG. 3 is a flowchart for explaining the operation in the first embodiment. FIG. 4 is a flowchart for explaining the operation shown in step ST10 of FIG. FIG. 5 is a schematic diagram for explaining a category query in the first embodiment. FIG. 6 is a schematic diagram for explaining a matrix representation of the category query in FIG. FIG. 7 is a schematic diagram showing a tree diagram included in the guideline A in the first embodiment. FIG. 8 is a schematic diagram for explaining guideline categories for each CQ included in guideline A in FIG. FIG. 9 is a schematic diagram for explaining the similarity between the category query in FIG. 5 and the guideline category in FIG. FIG. 10 is a schematic diagram for explaining a guideline extended category query in the first embodiment. FIG. 11 is a schematic diagram showing an importance table for explaining a guideline extended category query in the first embodiment. FIG. 12 is a schematic diagram for explaining calculation of the similarity in the first embodiment. FIG. 13 is a schematic diagram for explaining the importance of the guideline extended category query of FIG. FIG. 14 is a schematic diagram for explaining the expansion using the guideline expansion category query. FIG. 15 is a schematic diagram illustrating an example of a display screen according to the first embodiment. FIG. 16 is a schematic diagram showing another example of the display screen in the first embodiment. FIG. 17 is a schematic diagram showing yet another example of the display screen in the first embodiment. FIG. 18 is a schematic diagram showing yet another example of the display screen in the first embodiment. FIG. 19 is a schematic diagram for explaining a modified example of the first embodiment. FIG. 20 is a schematic diagram for explaining another modified example of the first embodiment. FIG. 21 is a schematic diagram for explaining yet another modified example of the first embodiment. FIG. 22 is a block diagram showing the configuration of a medical assistance device according to the second embodiment. FIG. 23 is a flowchart for explaining the operation in the second embodiment. FIG. 24 is a schematic diagram for explaining the search results of each guideline in the second embodiment. FIG. 25 is a block diagram showing the configuration of a medical assistance device according to the third embodiment. FIG. 26 is a flowchart for explaining the operation in the third embodiment.

Each embodiment will be explained below using the drawings. In the following explanation, guidelines will be used as an example of literature information related to medical treatment, among which guidelines, papers, case studies, etc. Accordingly, as an example of each itemized description section within a guideline (literature information), the section divided into the smallest items or clinical questions (CQs) will be used as an example of the section divided into clinical questions (CQs).

First Embodiment
FIG. 1 is a block diagram showing a medical support device and its peripheral configuration according to a first embodiment. The medical support device 1 shown in FIG. 1 is, for example, a device capable of comprehensively observing medical information. The medical support device 1 is equipped with, for example, an integrated viewer. The integrated viewer is an application that comprehensively presents medical information to a user. The integrated viewer may be implemented in any form, such as a web application, a fat client application, or a thin client application. The medical support device 1 is communicably connected to a hospital information system (HIS) 2, a radiology information system (RIS) 3, a medical image diagnostic device 4, a medical image management system (PACS: Picture Archiving and Communication Systems) 5, and a data warehouse (DWH: Data Ware House) 6 via an intra-hospital network such as a local area network (LAN).

In Figure 1, HIS2 includes, for example, an electronic medical record system that manages information related to electronic medical records. Information related to electronic medical records includes, for example, patient information and multiple pieces of medical information. Patient information is information specific to a patient, and includes, for example, a patient ID, patient name, gender, and age.

The multiple pieces of medical information are associated with a patient ID and represent information that medical professionals have learned about a patient's physical condition, medical condition, treatment, etc. during the course of medical care. Each piece of medical information individually contains various types of information, such as image information, examination history information, electrocardiogram information, vital sign information, medication history information, report information, medical chart information, and nursing record information. The various pieces of information within the medical information are distinguishable by data type. Similarly, the various pieces of information contained in each of the image information, examination history information, electrocardiogram information, vital sign information, medication history information, report information, medical chart information, and nursing record information are also distinguishable by data type. Image information is, for example, information indicating the location of medical images obtained by photographing a patient. Image information includes, for example, information indicating the location of medical image files (described below) generated by the medical image diagnostic device 4 as a result of an examination. Examination history information is, for example, information indicating the history of test results obtained as a result of specimen testing, bacterial testing, etc., performed on a patient. Electrocardiogram information is, for example, information related to an electrocardiogram waveform measured from a patient. Vital sign information is, for example, basic information related to the patient's life. Vital sign information includes, for example, pulse rate, respiratory rate, oxygen concentration, body temperature, blood pressure, and level of consciousness. Medication history information is, for example, information indicating the history of the amount of medication administered to a patient. Report information is, for example, information compiled by a radiologist in the radiology department who interprets medical images such as X-ray images, CT images, MRI images, and ultrasound images in response to an examination request from a medical doctor in a clinical department, summarizing the patient's condition and illness. Report information includes, for example, interpretation report information representing an interpretation report created by the radiologist with reference to medical image files stored in the PACS 5. Note that, because report information is generally stored in the PACS 5, the electronic medical record system can display the report information by retrieving it from the PACS 5.

Medical record information is, for example, information entered into an electronic medical record by a doctor or other person. Medical record information includes, for example, medical records at the time of hospitalization, the patient's medical history, and medication prescription history.

Nursing record information is, for example, information entered into an electronic medical record by a nurse or other person. Nursing record information includes nursing records from the time of hospitalization, etc.

Information related to the electronic medical record also includes, for example, examination implementation information. The examination implementation information is generated by the medical imaging diagnostic device 4 that performed the examination in accordance with the examination order information. The examination implementation information is information that represents the examination performed by the medical imaging diagnostic device 4. The examination implementation information includes the order number, examination UID (Unique ID), patient ID, modality type, imaging region, and imaging conditions. The examination UID is an identifier that can uniquely identify the examination. The modality type represents the modality used for imaging. Examples of modality types include "X-ray computed tomography device," "X-ray diagnostic device," "magnetic resonance imaging device," and "ultrasound diagnostic device." The imaging region corresponds to the examination region included in the examination order information. Examples of imaging regions include the abdomen, brain, and chest. Imaging conditions include the body position, imaging direction, and whether or not a contrast agent was used.

The HIS2 also includes, for example, an ordering system that manages reservation information, order information, etc. Note that the HIS2 may also be configured such that the electronic medical record system includes an ordering system.

Appointment information includes, for example, information about consultation appointments and examination appointments. Information about consultation appointments includes, for example, the consultation date, consultation time, reception number, requesting physician, and requesting department. Information about examination appointments includes, for example, the examination date, examination time, and reception number. Order information is, for example, information about orders requested by medical doctors, etc., such as order information for imaging tests, specimen tests, physiological tests, prescriptions, and medications. If the order information is examination order information requesting an imaging test, the examination order information includes, for example, an order number that can identify the test, a patient ID, examination type, examination area, and requester information. The order number is a number issued when examination order information is entered and is an identifier that uniquely identifies examination order information within, for example, a single hospital. Examination types include X-ray examinations, CT (Computed Tomography) examinations, MR (Magnetic Resonance) examinations, and RI (Radio Isotope) examinations. Examination areas include, for example, the abdomen, brain, and chest. Requester information includes the name of the medical department and the name of the doctor in charge. Information regarding test reservations is linked to order information.

RIS3 is a system that manages examination reservation information related to radiological examination work. RIS3 adds various setting information to examination order information entered by medical doctors in an order system included in HIS2, for example, and accumulates the information, managing the accumulated information as examination reservation information. RIS3 may also add various setting information to examination order information using an irradiation record that records various setting information set in the medical image diagnostic device 4 during past examinations. RIS3 sends examination orders to the medical image diagnostic device 4 in accordance with the examination reservation information. RIS3 also sends examination implementation information generated by the medical image diagnostic device 4 as a result of the examination being performed to the electronic medical record system included in HIS2.

The medical imaging diagnostic device 4 is a device that performs examinations by taking images of a patient. Examples of the medical imaging diagnostic device 4 include X-ray computed tomography devices, X-ray diagnostic devices, magnetic resonance imaging devices, nuclear medicine diagnostic devices, and ultrasound diagnostic devices. The medical imaging diagnostic device 4 performs examinations based on examination reservation information transmitted from the RIS 3, for example. The medical imaging diagnostic device 4 generates examination implementation information and transmits it to the RIS 3.

The medical image diagnostic device 4 also generates medical image data by conducting examinations. Examples of medical image data include X-ray CT image data, X-ray image data, MRI image data, nuclear medicine image data, and ultrasound image data. The medical image diagnostic device 4 generates a medical image file by converting the generated medical image data into a format that complies with the DICOM (Digital Imaging and Communication in Medicine) standard, for example. The medical image file is, for example, a file in a format that complies with the DICOM standard. The medical image diagnostic device 4 transmits the generated medical image file to the PACS 5.

PACS 5 is a system that manages various medical image files. For example, PACS 5 stores medical image files transmitted from medical image diagnostic apparatus 4. PACS 5 may also store report information attached to medical image files or report information for examinations related to multiple medical image files.

DWH6 is a database system that collectively stores medical information (medical big data) generated by medical and nursing care institutions, and also stores multiple literature information related to medical treatments provided by medical and nursing care institutions. The multiple literature information includes at least one guideline. DWH6 is realized, for example, by a general server device. As shown in Figure 1, DWH6 has, for example, a processing circuit 61, memory 62, and communication interface 63. The processing circuit 61, memory 62, and communication interface 63 are connected to each other so that they can communicate with each other, for example, via a bus.

The processing circuitry 61 is a processor that functions as the core of the DWH 6. The processing circuitry 61 executes programs stored in the memory 62 and other devices to realize the functions corresponding to the programs. These functions include, for example, collecting desired information from the HIS 2, the RIS 3, the medical imaging diagnostic device 4, and the PACS 5, and storing the collected information in the memory 62, as appropriate. This allows, for example, electronic medical record information and literature information related to medical treatment to be collected from the HIS 2, examination appointment information to be collected from the RIS 3, and medical image files to be collected from the medical imaging diagnostic device 4 or the medical image management system 5. Literature information related to medical treatment includes guidelines, papers, case descriptions, and the like, each of which describes information according to medical specialty, disease, etc. Below, as described above, guidelines are used as an example of literature information. Furthermore, for example, medical information collected from the HIS 2 is stored in the memory 62 according to preset rules. The preset rules are, for example, sequences that use event dates and times associated with medical events such as outpatient visits, surgeries, imaging tests, specimen tests, bacteriological tests, electrocardiogram measurements, vital sign measurements, medication administration, report creation, and medical record entries for each patient. Event dates and times are, for example, the dates and times when medical events occur or when medical events are scheduled. Event dates and times include, for example, the dates and times when imaging tests are performed, the dates and times when specimen tests are performed, the dates and times when electrocardiogram waveforms are measured, the dates and times when vital signs are measured, the dates and times when medications are administered, the dates and times when reports are created, and the dates and times when medical records are entered. As a result, the collected medical information is stored in memory 62 in order of event dates and times, for example.

Memory 62 is a storage device such as an HDD (hard disk drive), SSD (solid state drive), or integrated circuit storage device that stores various information. Memory 62 may also be a drive that reads and writes various information to and from portable storage media such as a CD-ROM drive, DVD drive, or flash memory. Memory 62 stores control programs and the like that cause processing circuit 61 to perform various functions, such as collecting desired information and storing the collected information in memory 62. These programs may be distributed and stored in a non-transitory storage medium, and then read from the non-transitory storage medium and installed in memory 62.

The communication interface 63 communicates data with the medical support device 1, HIS 2, RIS 3, medical image diagnostic device 4, and PACS 5, which are connected via the hospital network. Any standard may be used for communication with the medical support device 1, HIS 2, RIS 3, medical image diagnostic device 4, and PACS 5, such as Hearth Level 7 (HL7), DICOM, or both.

Next, details of the medical assistance device 1 according to this embodiment will be described with reference to Figure 2. Figure 2 is a block diagram showing the functional configuration of the medical assistance device 1 shown in Figure 1.

The medical assistance device 1 shown in FIG. 2 includes a memory 11, an input interface 12, a display 13, a communication interface 14, and a processing circuit 15. The memory 11, the input interface 12, the display 13, the communication interface 14, and the processing circuit 15 are connected to each other so that they can communicate with each other, for example, via a bus.

Here, memory 11 is composed of memories that record electrical information, such as ROM (Read Only Memory), RAM (Random Access Memory), HDD (Hardware Disk Drive), and image memory, as well as peripheral circuits associated with these memories, such as a memory controller and memory interface. Memory 11 stores, for example, various programs, such as the medical information processing program of this medical support device, and various data, such as information about electronic medical records obtained from DWH6, patient medical information, medical guidelines, guideline categories for each term group corresponding to each clinical question (CQ) in the guidelines, various tables, data in progress, and data after processing. Note that information about electronic medical records includes, for example, patient information including a patient ID and medical information associated with the patient ID. Details of the patient information and medical information are as described above.

The input interface 12 may be implemented by a trackball, switch buttons, mouse, keyboard, touchpad (or trackpad) for inputting various instructions, commands, information, selections, and settings from the operator (user) into the medical assistance device, or a touch panel display (or touch screen) that integrates a display screen and a touchpad. The input interface 12 is connected to the processing circuitry 15 and converts input operations received from the user into electrical signals and outputs them to the processing circuitry 15. In this case, the input interface 12 may display a user interface (GUI: Graphical User Interface) on the display 13, allowing the user to input various instructions using physical operation components such as a mouse or keyboard. Note that, in this specification, the input interface 12 is not limited to those equipped with physical operation components. For example, an electrical signal processing circuit that receives electrical signals corresponding to input operations from an external input device provided separately from the device and outputs these electrical signals to the processing circuitry 15 is also included as an example of the input interface 12. In the following description, "operation of the input interface 12 by the user" is also referred to as "user operation."

The display 13 is composed of a display main body that displays literature information, etc., an internal circuit that supplies display signals to the display main body, and peripheral circuits such as connectors and cables that connect the display to the internal circuitry. The display 13 can display any data as needed, for example, a category query indicating a group of medical terms contained in a patient's medical information, guideline categories, and literature information search results. The display 13 is an example of a display unit.

The communication interface 14 is a circuit that connects the medical assistance device 1 to a network and enables communication with other devices. For example, a network interface card (NIC) can be used as the communication interface 14. In the following description, the involvement of the communication interface 14 in communication between the medical assistance device 1 and other devices will be omitted.

The processing circuitry 15 reads the medical assistance program stored in the memory 11 based on instructions input by the user via the input interface 12 and controls the medical assistance device 1 in accordance with these instructions. For example, the processing circuitry 15 is a processor that implements each function of the medical assistance device 1 in accordance with the medical assistance program read from the memory 11. Examples of each function include a category query creation function 15a, a guideline search function 15b, a query expansion function 15c, and a display control function 15d. Each function may be distributed across multiple processors as appropriate. Alternatively, each function or part of each function may be executed by another device as appropriate. For example, among the functions, the category query creation function 15a may be executed by another device (not shown). In other words, the category query creation function 15a is an optional additional feature that is not necessarily required for the medical assistance device 1 and may be omitted from the medical assistance device 1.

Next, the functions of the category query creation function 15a, guideline search function 15b, query expansion function 15c, and display control function 15d will be described in order. However, the allocation of functions described below is for convenience and can be changed as appropriate. This is because even if a process that is assigned to one function is assigned to another function, the processing circuitry 15 still executes that process. Note that the ability to change the allocation of functions also applies to the following embodiments and variants.

The category query creation function 15a acquires a patient's medical information from the DWH 6 and creates a category query from a group of medical terms contained in the medical information. Here, a category query is a query that expresses the relationships between words grouped by medical meaning, stratified according to upper and lower levels. Specifically, for example, the category query creation function 15a extracts each medical term contained in the patient's medical information by classification category related to medical treatment, and creates a category query from the group of medical terms containing medical terms for each classification category. Note that a classification category is each of multiple classification categories with a hierarchical relationship. Examples of classification categories that can be used as appropriate include disease name, treatment phase, disease type, treatment method, and symptoms. For example, in the case of cancer treatment, guidelines are categorized from top to bottom in the following order: treatment phase (diagnosis or treatment), subtype, disease stage (or metastatic site), treatment method, and clinical question (CQ). Therefore, the same medical term may belong to different classification categories, and a simple term search alone may not link to the page the doctor wants to refer to. Taking this into consideration, as described above, medical terminology groups for each classification category are created as category queries. The category query creation function 15a is an example of a medical terminology group creation unit.

The guideline search function 15b extracts, from among multiple guidelines related to medical treatment, a first guideline related to a group of medical terms (category query) included in the patient's medical information. Furthermore, the guideline search function 15b extracts one of the first guideline categories from among the first term groups (first guideline categories) corresponding to each clinical question (CQ) in the first guideline based on a first relevance with the category query. The first guideline categories are created in advance as first term groups containing first medical terms for each classification category by extracting each first medical term included in each clinical question (CQ) in the first guideline by classification category related to medical treatment. The classification categories are the same as the classification categories of the category query. Therefore, for example, if the first term group includes a first treatment method term representing a first treatment method, the first treatment method term is extracted so as to be included in the classification category indicating the treatment method. The first relevance is, for example, a value based on a first similarity between each first guideline category and the category query.

Similarly, the guideline search function 15b extracts one of the second guideline categories from each second term group (second guideline category) corresponding to each clinical question (CQ) in a second guideline different from the first guideline, based on the second relevance between the extracted first guideline category and the guideline extension category obtained from the category query. The second guideline category is created in advance as a second term group containing second medical terms by classification category related to medical treatment, by extracting each second medical term contained in each clinical question (CQ) in the second guideline by classification category related to medical treatment. The classification category is the same as the classification category of the category query. Therefore, for example, if the second term group includes a second treatment method term representing a second treatment method different from the first treatment method, the second treatment method term is extracted so as to be included in the classification category indicating the treatment method. The second relevance is, for example, a value based on the second similarity between each second guideline category and the extracted first guideline category and category query. The guideline search function 15b is an example of a first extraction unit, a second extraction unit, a third extraction unit, a first creation unit, and a second creation unit.

The query expansion function 15c creates a guideline extension category from the extracted first guideline category and category query. The guideline extension category is used by the guideline search function 15b to search for (extract) other guideline categories.

The display control function 15d outputs the first guideline and the second guideline based on the extracted first guideline category and the extracted second guideline category. For example, the display control function 15d causes the display 13 to display and output the clinical questions (CQs) in the first guideline that correspond to the extracted first guideline category. The display control function 15d also causes the display 13 to display and output the clinical questions (CQs) in the second guideline that correspond to the extracted second guideline category.

In addition, the display control function 15d displays and outputs on the display 13 clinical questions (CQs) in the first guideline that include the first treatment method term and correspond to the extracted first guideline category. If the first treatment method indicated by the first treatment method term being displayed and output causes a symptom in the medical information, the display control function 15d displays and outputs on the display 13 clinical questions (CQs) in the second guideline that include the second treatment method term indicating a second treatment method that improves the symptom and corresponds to the extracted second guideline category. The display control function 15d is an example of an output unit.

Next, the operation of a medical information processing system equipped with the medical support device configured as described above will be explained using the flowcharts in Figures 3 and 4 and the schematic diagrams in Figures 5 to 18.

Now, assume that the DWH 6 stores patient information, patient clinical information, each guideline, and each guideline category in memory 62. In this state, step ST10 is started, in which a category query is created from the patient clinical information. Step ST10 includes steps ST11 to ST17.

(Step ST10)
The processing circuitry 15 of the medical assistance device 1 acquires the patient's medical information from the DWH 6 based on the patient ID of the patient (step ST11), and stores the acquired medical information in the memory 11.

The processing circuit 15 determines whether or not a diagnosis is included in the patient's medical information (step ST12). If a diagnosis is not included, the patient has not yet been diagnosed, so the processing circuit 15 extracts symptoms from the medical information (step ST13) and proceeds to step ST17.

Furthermore, if the result of the determination in step ST12 indicates that a diagnosis has been made, the processing circuitry 15 determines that the diagnosis has been made and therefore extracts the subtype and stage of the disease as the disease type from the medical information (step ST14). Furthermore, the processing circuitry 15 extracts the treatment history as the treatment history from the medical information (step ST15). Furthermore, the processing circuitry 15 extracts symptoms from the medical information (step ST16).

In step ST13 or steps ST14 to ST16, the processing circuitry 15 extracts each medical term contained in the medical information by classification category related to medical treatment. The definitions of the classification categories are determined by the chapter divisions of disease-related guidelines and by medical professionals.

Then, the processing circuit 15 creates a category query q1, which is a group of medical terms containing medical terms classified by hierarchical category, as shown in Figure 5 (step ST17).

This category query q1 includes the disease name L1 "lung cancer," the treatment phase L2 "post-diagnosis," the disease type L3 "non-small cell lung cancer," "stage III," and "suspected bone metastasis," the treatment method L4 "radiation therapy administered," and the symptom L5 "numbness in hands and feet." For this category query q1, the processing circuit 15 converts the medical terminology group included in the classification category into a matrix representation Mq1 using a dictionary-based one-hot representation, as shown in FIG. 6. This completes step ST10, which consists of steps ST11 to ST17.

(Step ST20)
In step ST20, the processing circuit 15 searches for a guideline category based on the category query. In the following description, the first guideline is referred to as guideline A, the second guideline is referred to as guideline B, and the third guideline is referred to as guideline C.

First, the processing circuit 15 extracts guideline A related to the category query included in the patient's medical information from among multiple guidelines related to medical treatment.

As shown in Figures 7 and 8, the guideline categories for guideline A were created in advance by using the chapter structure and tree diagram F1 of guideline A7 to extract each first medical term included in each clinical question (CQ) into classification categories related to medical treatment. In Figure 8, guideline category CA1 was created corresponding to clinical question CQ19 in guideline A7, and guideline category CA2 was created corresponding to clinical question CQ21 in guideline A7. Guideline categories CA1 and CA2 have a structure similar to the category query q1 described above, but unlike category query q1, related medical terms are connected by edges. Examples of related medical terms include medical terms included in chapters related to higher-level medical terms in the chapter structure. Related items include medical terms that appear in the same paragraph. Medical terms may also be extracted using a dictionary that lists the terms to be extracted.

Next, the processing circuit 15 extracts one of the guideline categories CA1, CA2, ... corresponding to each clinical question (CQ) in the guideline A7 based on the first relevance with the category query q1. Here, the first relevance is the similarity between each guideline category CA1, CA2, ... and the category query q1. For example, as shown in FIG. 9, the processing circuit 15 calculates the similarity between each guideline category CA1, CA2, ... and the category query q1, and extracts the guideline category CA2 with the highest similarity. Note that the similarity is calculated, for example, as the inner product of the matrix representation Mq1 of the category query q1 and the matrix representation of each guideline category CA1, CA2, .... This completes step ST20.

(Step ST30)
In step ST30, the processing circuitry 15 expands the searched guideline category CA2 based on the category query q1. For example, as shown in Fig. 10, the processing circuitry 15 combines the category query q1 with the searched guideline category CA2 to create a new guideline expansion category query ECA2. Note that in Fig. 10, the guideline expansion category query ECA2 changes the attribute of medical terms in the guideline category CA2 that are supported by fact (category query q1) from hypothesis (dotted frame) to fact (solid frame).

Specifically, such a guideline expanded category query ECA2 is created as follows. Specifically, the processing circuit 15 performs a logical OR operation on the medical terms included in the category query q1 and the guideline category CA2 in the search results. Information such as edges is retained. Furthermore, the processing circuit 15 increases the weighting of medical terms with high overall importance based on the importance table T1 shown in FIG. 11. The importance table T1 is used to calculate an index of overall importance from four indices related to the importance of medical terms. The four indices are "fact/hypothesis," "presence or absence of an edge with fact," "term importance," and "category query-specific importance." The "fact/hypothesis" indicator indicates the importance of terms recognized as facts from patient data (terms in the category query q1). In other words, since terms listed in the guidelines include possibilities and options, this indicator distinguishes them as hypotheses. Furthermore, the "presence or absence of an edge with fact" indicator indicates the importance of information that is strongly related to terms recognized as facts (i.e., information that has an edge connected to a term recognized as fact). The indicator "term importance" is an importance that is predefined based on prior knowledge of doctors and guidelines. The indicator "category query-specific importance" is an importance determined for each patient for the extracted words and layers using data mining or other methods based on the patient's medical information. For example, the category query-specific importance may be determined based on the frequency of term appearance. The category query-specific importance may also be determined based on a rule base depending on the patient's condition. For example, the rule base may indicate that administering drug X results in a high importance of "drug rash."

In Figure 11, the value of the indicator "fact/hypothesis" is "fact" if it matches a term in category query q1, and "hypothesis" otherwise. The value of the indicator "presence or absence of edge with fact" is "1" if there is an edge to the term that is "fact," and "0" otherwise. The value of the indicator "overall importance" is calculated by weighting the values of the other indicators. For the indicator "fact/hypothesis," the weighting value is 0.3 for "fact" and 0 for "hypothesis." For the indicator "presence or absence of edge with fact," the weighting value is 0.1 for "edge (1)" and 0 for "no edge (0)." The weighting value is 1 for both the indicator "term importance" and the indicator "category query-specific importance" (importance is used as is). Therefore, for example, the value of the indicator "overall importance" for the term "non-small cell lung cancer" is calculated as 0.3 (fact) + 0.1 (edge) + 0.4 + 0.4 = 1.2.

The processing circuit 15 converts this guideline extended category query ECA2 into a matrix representation using the overall importance from the importance table T1. This completes step ST30.

(Step ST40)
In step ST40, the processing circuit 15 searches for other guideline categories based on the guideline extension category.

First, the processing circuit 15 selects another guideline. As a selection method, for example, the following methods (1) to (4) can be used as appropriate.

(1) A method predefined by a doctor.

(2) A way to group highly related guidelines.

(3) A method for automatically selecting highly relevant guidelines based on the patient's condition.

(4) A method for calculating the similarity between a patient's medical information and the entire guideline using medical terms that appear in the guideline and the patient's medical information.

In addition, when searching for guidelines, you may search for the same guideline repeatedly. This is because repeatedly searching for the same guideline based on guideline extension categories using other guidelines allows you to perform a more relevant search.

Here, assume that the processing circuitry 15 selects guideline B, which is different from guideline A. As shown in FIG. 12, the processing circuitry 15 extracts one of the guideline categories CB1, ... corresponding to each clinical question (CQ) in guideline B8 based on the second relevance with the guideline extended category query ECA2. The second relevance is the similarity between each guideline category CB1, ... and the guideline extended category query ECA2.

For example, as described above, the processing circuit 15 calculates the similarity between each guideline category CB1, ... and the guideline extension category query ECA2, and extracts the guideline category CB1 with the highest similarity. Note that the similarity is calculated as the inner product of the matrix representation M_ECA2 of the guideline extension category query ECA2 and the matrix representation of each guideline category CB1, ..., as shown in FIG. 13, for example. This completes step ST40.

(Step ST50)
In step ST50, the processing circuit 15 expands another guideline category CB1 based on the guideline expansion category query ECA2, as shown in FIG.

(Step ST60)
In step ST60, the processing circuit 15 determines whether or not the process has ended, and if not, returns to step ST40 to continue the process, whereas if the result of the determination is that the process has ended, the process proceeds to step ST70.

(Step ST70)
In step ST70, the processing circuitry 15 displays guidelines A7, B8, C, ... on the display 13 based on the searched guideline categories CA2, CB1, .... For example, the processing circuitry 15 displays a clinical question (CQ) in guideline A corresponding to the searched guideline category CA2 on the display 13. Also, for example, the processing circuitry 15 displays a clinical question (CQ) in guideline B corresponding to the searched guideline category CB1 on the display 13. Specifically, for example, the processing circuitry 15 displays a clinical question (CQ) in guideline A7 corresponding to the extracted guideline category CA2, which includes a first treatment method term, on the display 13. When the first treatment method indicated by the first treatment method term being displayed causes a symptom in the medical information, the processing circuit 15 displays and outputs on the display 13 a clinical question (CQ) in the guideline B corresponding to the extracted guideline category CB1 that includes a second treatment method term indicating a second treatment method that improves the symptom.

Figures 15 to 18 are schematic diagrams showing examples of display screens displayed on the display 13. As shown in Figure 15, one display screen displays a medical record entry corresponding to the patient's medical information in tab Tb1, guideline A7 in tab Tb2 of guideline candidates, and guideline B8 in tab Tb3 of guideline candidates.

In tab Tb1, medical terms of high importance are highlighted HL1. Furthermore, operating the Expand Guideline button b1 expands the guideline A7 related to the highlighted medical term.

Similarly, in tab Tb2, medical terms with high importance are highlighted with highlights HL1 and HL2. Operating the guideline expansion button b2 expands guidelines B8 related to the highlighted medical term. The expansion buttons b1 to b3 may display a list of highly relevant guidelines using a pull-down menu or the like, allowing selection. As shown in tabs Tb1 and Tb2, clinical questions (CQs) from search results may be displayed with highlights HL1 and HL2 using a tree diagram or the like. Highlights HL1, HL2, HL3, etc. are each represented by a different color, and highlights of the same color (same symbol) indicate information in the same category.

Next, another display screen, as shown in Figure 16, displays clinical questions (CQs) from related guidelines based on treatment options.

For example, suppose a clinical question (CQ) is displayed in guideline A7 that includes the treatment method options "surgery" and "radiotherapy." Here, when the option "radiotherapy" in guideline A7 is selected, a clinical question (CQ) is displayed that includes the treatment method option "external beam radiation therapy" in guideline B8, which is related to "radiotherapy." Similarly, when the option "external beam radiation therapy" in guideline B8 is selected, a clinical question (CQ) is displayed that includes the treatment method option "radiotherapy" in guideline C9, which is related to "external beam radiation therapy."

Next, another display screen, as shown in Figure 17, displays clinical questions (CQs) from the guidelines related to the possible symptoms.

For example, from "Platinum Therapy" and "Anticancer Drug A" in the patient information, "Neutropenia," "Thrombocytopenia," and "Interstitial Pneumonia" are displayed as possible symptoms in the chemotherapy CQs in Guideline A7. Here, when the displayed symptom "Neutropenia" is selected, other clinical questions (CQs) in Guideline A7 related to that symptom are displayed.

Next, as shown in Figure 18, another display screen displays related guidelines based on the guideline content and differences between the patient's condition and treatment method.

For example, if the patient's condition is "BMA administration" and "Ca decrease," and guideline A lists the treatment method as "bone-modifying drug" and the adverse event as "hypocalcemia," then the clinical questions (CQs) in the related guideline B, including "Radiation therapy prevents pathological fractures and bone marrow compression," will be displayed based on "BMA" and "hypocalcemia" in the relevant guideline B.

Specifically, for example, the processing circuitry 15 displays and outputs on the display 13 a clinical question (CQ) in guideline A7 corresponding to the extracted guideline category CA2, which includes the first treatment method term "bone-modifying drug." If the first treatment method indicated by the first treatment method term being displayed and output causes the symptom "Ca decrease" in the medical information, the processing circuitry 15 displays and outputs on the display 13 a clinical question (CQ) in guideline B8 corresponding to the extracted guideline category CB1, which includes the second treatment method term "radiation therapy," which indicates a second treatment method that improves the symptom.

As described above, according to the first embodiment, a first guideline related to a group of medical terms (category query) included in the patient's medical information is extracted from among multiple guidelines related to medical treatment. Furthermore, from among the first groups of terms (first guideline categories) corresponding to each clinical question (CQ) in the first guideline, any first guideline category is extracted based on a first degree of relevance with the category query. Furthermore, from among the second groups of terms (second guideline categories) corresponding to each clinical question (CQ) in a second guideline different from the first guideline, any second guideline category is extracted based on a second degree of relevance with the extracted first guideline category and a guideline extension category obtained from the category query. Furthermore, the first guideline and second guideline are output based on the extracted first guideline category and the extracted second guideline category. Therefore, it is possible to present multiple pieces of literature information that may be relevant to the patient while reducing the effort required to refer to literature information related to medical treatment.

Additionally, conventional systems are unable to reference knowledge from other related fields of expertise via guidelines. In contrast, according to the first embodiment, by starting from the patient's condition and expanding multiple guidelines in a chain, it is possible to integrate and display information from multiple guidelines related to the patient. This also broadens treatment options. It also allows for early consultation with doctors from other departments.

Furthermore, according to the first embodiment, clinical questions (CQs) in the first guideline corresponding to the extracted first guideline category are displayed on the display 13. Furthermore, the display control function 15d displays clinical questions (CQs) in the second guideline corresponding to the extracted second guideline category on the display 13.
In this way, the clinical questions (CQs) in the guidelines are displayed, so that the minimum items that one wishes to refer to can be displayed.

Furthermore, according to the first embodiment, the first guideline category includes a first treatment method term representing a first treatment method, and the second guideline category includes a second treatment method term representing a second treatment method different from the first treatment method. Furthermore, clinical questions (CQs) in the first guideline that include the first treatment method term and correspond to the extracted first guideline category are displayed and output on the display 13. Furthermore, if the first treatment method indicated by the first treatment method term being displayed and output causes a symptom in the medical information, clinical questions (CQs) in the second guideline that include a second treatment method term indicating a second treatment method that improves the symptom and correspond to the extracted second guideline category are displayed and output on the display 13. Therefore, if a first treatment method that causes a symptom is displayed, a second treatment method that improves the symptom can be displayed.

Furthermore, according to the first embodiment, by extracting each medical term contained in a patient's medical information by classification category related to medical treatment, a group of medical terms containing medical terms by classification category is created as a category query. This makes it possible to create a category query suitable for searching for guidelines from the patient's medical information.

Furthermore, according to the first embodiment, first medical terms included in each clinical question (CQ) in the first guideline are extracted by classification category related to medical treatment, and a first guideline category containing first medical terms by classification category is created. This makes it possible to create a first guideline category from the first guideline that is suitable for searching each clinical question (CQ). This also applies to a second guideline that is different from the first guideline.

Furthermore, according to the first embodiment, the classification category is each of a plurality of classification categories that have a hierarchical relationship. Therefore, for example, even if the same medical term belongs to different classification categories and a simple term search alone does not yield the results a doctor wants to see, a term search that takes classification categories into consideration can be expected to yield the results the doctor wants to see.

Furthermore, according to the first embodiment, the first relevance is a value based on the first similarity between each first guideline category and the category query. The second relevance is a value based on the second similarity between each second guideline category and the extracted first guideline category and category query. In this case, the similarity can be easily obtained by calculating the inner product of the matrix representation of the guideline category and the matrix representation of the category query.

Furthermore, according to the first embodiment, the multiple pieces of literature information include at least one guideline, and the itemized description portion of the guideline is a portion that is described by dividing it into the smallest items or clinical questions (CQs). Therefore, depending on the search results, the clinical questions (CQs) or smallest items portion can be displayed.

[Modification of the first embodiment]
The first embodiment as described above may be implemented as shown in the following modified examples.

[1] The category query creation function 15a may predefine the medical term groups and classification categories to be extracted. Here, the medical term groups to be extracted (a dictionary of target words) may be predefined using prior knowledge of doctors and guidelines. Alternatively, medically important word groups may be extracted as medical term groups using data mining that utilizes natural language processing. Alternatively, TF-IDF (a method for calculating importance based on word distribution) may be used, or important words may be extracted using machine learning, etc.

The classification categories (word classifications) to be extracted may be determined by using machine learning or other methods to classify medical terms into topics, and defining these topics as classification categories. Alternatively, classification categories may be defined using topic analysis and clustering techniques.

[2] The category query creation function 15a may define in advance the importance of the extracted medical terms and classification category hierarchies, and use this in query searches. For example, as shown in FIG. 19, a weight matrix indicating the importance for each medical term or classification category L1 to L5 may be defined, and this may be integrated into the matrix representation Mq1 of the category query q1 and used in query searches. Here, the importance may be defined in advance based on prior knowledge of doctors or guidelines, or the importance of medical terms and hierarchies may be defined from intermediate data of the machine learning method used in [1] above. The intermediate data of the machine learning method may be, for example, neural network parameters (weights, biases). Furthermore, the importance may be adjusted individually, or the overall importance may be raised.

[3] The category query creation function 15a may use importance specific to the category query. Specifically, the importance of medical terms may be determined for each patient using data mining or other methods from the patient's medical information for the extracted medical term group or classification category hierarchy. For example, the importance may be determined based on the frequency of appearance of the medical term. Alternatively, the importance may be determined based on a rule-based system, such as "administering drug X → 'drug rash' has a high importance" depending on the patient's condition.

"4" The category query creation function 15a may create multiple category queries with different structures. For example, multiple category queries may be created depending on the type of search target. Specifically, a category query may be created for each disease (when a patient suffers from multiple diseases). For example, in the case of cancer, a category query for specimen test results such as tumor markers may be created. Similarly, in the case of heart disease, a category query for vital signs information may be created. Also, a category query may be created for each treatment method (in the case of combination therapy). For example, in the case of chemotherapy, a category query with an additional layer of drug information may be created. Furthermore, in the case of radiation therapy, a category query with an additional layer of radiation therapy planning may be created.

Furthermore, this allows the processing circuitry 15 to perform a search using multiple category queries. In this case, any search method, such as AND search, OR search, or ranking search, can be used as appropriate.

[5] The guideline search function 15b may rank search results taking into consideration the recommendation level of the clinical question (CQ). Specifically, as shown in FIG. 20, related guideline categories CA2 may be searched for using the product of similarity and recommendation level. In other words, the first relevance level may be the product of the first similarity and the recommendation level described for the clinical question (CQ) corresponding to each first guideline category.

[6] As shown in FIG. 21, the category query creation function 15a may create a category query by specifying a search query by the user. For example, a guideline search tab Tb0 is displayed on the display 13. The tab Tb0 includes a main guideline list DL1, a search guideline list DL2, a search axis DL3, a search term input field IA1, and a guideline search button b0. Therefore, when the main guideline, search guideline, search axis, search term, etc. are specified in response to user operation, the processing circuit 15 creates a category query based on this specification.

Second Embodiment
Next, a medical assistance device according to a second embodiment will be described with reference to Fig. 22. In the following description, elements that are substantially the same as those in the above-mentioned drawings will be assigned the same reference numerals and detailed description thereof will be omitted, and different elements will be mainly described.

The second embodiment takes into consideration situations where you want to search not only for highly relevant items, but also to prevent oversights, and searches through guidelines to find statements that could lead to serious situations such as aggravation of the condition.

Specifically, as shown in FIG. 22, the processing circuit 15 further includes an important category search function 15e. Furthermore, at least one of the first guideline categories CA1, CA2, ... and the second guideline categories CB1, ... includes important terms related to serious situations.

Here, the important category search function 15e searches for guideline categories that include important terms from the extracted first guideline category CA1 and second guideline category CB1. Note that the important category search function 15e is an example of a search unit.

As a result, the processing circuitry 15 preferentially outputs clinical questions (CQs) from the first guideline A and the second guideline B that correspond to the searched term group.

Other configurations are the same as in the first embodiment.

With the above configuration, as shown in the flowchart of Figure 23, before step ST20 of searching for guideline categories, the processing circuit 15 prepares guideline categories containing key terms in response to a user operation (step ST20_prep). Note that step ST20_prep may also be performed before step ST10.

In either case, after step ST20_prep, steps ST20 to ST60 are executed as described above.

After step ST60, in step ST70a, the processing circuit 15 searches for guideline categories that include key terms among the extracted first guideline category CA1 and second guideline category CB1. As a result of the search, for example, as shown in FIG. 24, the second guideline category CB1 that includes key term Tm3 was obtained. Furthermore, medical term Tm2 in the first guideline CA1 was obtained as a term related to key term Tm3, and medical term Tm1 in category query q1 was obtained as a term related to medical term Tm2.

In response to this, the processing circuitry 15 will prioritize output of clinical questions (CQs) corresponding to the searched term group from the first guideline A and the second guideline B. In this case, medical terms Tm1, Tm2, and Tm3 linked to key terms will be displayed with priority.

As described above, according to the second embodiment, at least one of the first guideline categories CA1, CA2, ... and the second guideline categories CB1, ... includes key terms related to serious situations. Furthermore, the extracted first guideline category CA1 and second guideline category CB1 are searched for guideline categories that include key terms. Furthermore, clinical questions (CQs) corresponding to the searched term group are preferentially output from the first guideline A and the second guideline B. Therefore, in addition to the effects of the first embodiment, it is possible to prevent overlooking key terms related to serious situations.

Third Embodiment
Next, a medical assistance device according to a third embodiment will be described with reference to FIG.

The third embodiment takes into account cases where a doctor wants to search for possible symptoms or treatments that are not included in the patient's medical information, and by analyzing the doctor's behavior, the category query is modified to reflect the information the doctor wants to refer to.

Specifically, as shown in FIG. 25, the processing circuit 15 further includes a behavioral analysis processing function 15f.

Here, the behavior analysis processing function 15f analyzes the operator's behavior while the guideline is being output, modifies the category query q1 to add medical terms corresponding to the analysis results, and re-executes the guideline search function 15b, query expansion function 15c, and display control function 15d in accordance with the modified category query q1. The behavior analysis processing function 15f is an example of a behavior analysis unit.

Other configurations are the same as those of the second embodiment. However, the third embodiment may also be applied to the first embodiment.

With the above configuration, as shown in the flowchart of Figure 23, after step ST70a in which the guidelines are displayed, the processing circuit 15 determines whether or not the user has finished depending on whether or not an end operation has been performed (step ST80), and if the user has not finished, analyzes the operator's behavior (step ST80).

The processing circuitry 15 also modifies the category query q1 to add medical terms that correspond to the analyzed results. Alternatively, the processing circuitry 15 may enhance the importance of terms or edges based on the search results selected by the doctor. In either case, the processing circuitry 15 re-executes the guideline search function 15b, the query expansion function 15c, and the display control function 15d in response to the modified category query q1.

Then, based on the revised category query q1, the processing from step ST20_prep onwards is executed in the same manner as described above.

As described above, according to the third embodiment, the operator's behavior is analyzed while the guidelines are being output, the category query q1 is modified to add medical terms corresponding to the analysis results, and the guideline search function 15b, query expansion function 15c, and display control function 15d are re-executed in accordance with the modified category query q1. Therefore, in addition to the effects of the second embodiment, the category query can be modified to relate to the information that the doctor wishes to refer to, taking into account cases where the doctor wishes to search for possible symptoms or treatments.

At least one of the embodiments described above makes it possible to present multiple pieces of literature information that may be relevant to the patient while reducing the effort required to refer to literature information related to medical treatment.

The term "processor" used in the above description refers to circuits such as a CPU (central processing unit), GPU (Graphics Processing Unit), or 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)). If the processor is a CPU, for example, the processor realizes its functions by reading and executing a program stored in a memory circuit. On the other hand, if the processor is an ASIC, for example, the program is not stored in a memory circuit, but the function is directly incorporated into the processor circuit as a logic circuit. Note that 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. Furthermore, multiple components in Figures 1, 2, 22, or 25 may be integrated into a single processor to realize its function.

While several embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These embodiments can be embodied in a variety of other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their variations are included within the scope of the invention and its equivalents as set forth in the claims, as well as within the scope and spirit of the invention.

REFERENCE SIGNS LIST 1 medical support device 11, 62 memory 12 input interface 13 display 14, 63 communication interface 15, 61 processing circuit 15a category query creation function 15b guideline search function 15c query expansion function 15d display control function 15e important category search function 15f behavior analysis processing function 2 HIS
3 RIS
4 Medical imaging diagnostic equipment 5 PACS
6 DWH

Claims (12)

a first extraction unit that extracts first literature information related to a group of medical terms included in the patient's medical information from a plurality of literature information related to medical treatment;
a second extraction unit that extracts one of the first term groups corresponding to each of the itemized description portions in the first literature information based on a first degree of association with the medical term group;
a third extraction unit that extracts any one of the second term groups corresponding to each of the itemized description portions in second literature information different from the first literature information based on a second degree of association between the extracted first term group and the medical term group;
an output unit that outputs the first literature information and the second literature information based on the extracted first term group and the extracted second term group. the output unit causes a display unit to display and output an itemized description portion in the first literature information corresponding to the extracted first term group, and causes a display unit to display and output an itemized description portion in the second literature information corresponding to the extracted second term group.
The medical support device according to claim 1. the first term group includes a first treatment method term representing a first treatment method;
the second term group includes a second treatment method term representing a second treatment method different from the first treatment method;
the output unit causes a display unit to display and output an itemized description portion in the first literature information that includes the first treatment method term and corresponds to the extracted first term group, and when a first treatment method indicated by the first treatment method term being displayed and output causes a symptom in the medical information, causes the display unit to display and output an itemized description portion in the second literature information that includes a second treatment method term indicating the second treatment method that improves the symptom and corresponds to the extracted second term group.
3. The medical support device according to claim 1 or 2. a medical term group creation unit that creates the medical term group including medical terms for each classification category by extracting each medical term included in the medical information for each classification category related to medical treatment;
The medical support device according to claim 1 , further comprising: a first creation unit that creates the first term group including the first medical terms by classification category related to medical treatment by extracting each first medical term included in each itemized description portion in the first literature information by classification category related to medical treatment;
The medical support device according to claim 1 , further comprising: a second creation unit that creates the second term group including the second medical terms by classification category related to medical treatment by extracting each second medical term included in each itemized description portion in the second literature information by classification category related to medical treatment;
The medical support device according to claim 1 , further comprising: The medical support device described in any one of claims 4 to 6, wherein the classification category is each of a plurality of classification categories having a hierarchical relationship. the first relevance is a value based on a first similarity between each of the first term groups and the medical term group;
8. The medical support device according to claim 1, wherein the second relevance is a value based on a second similarity between each of the second term groups and the extracted first term group and the medical term group. The medical support device of claim 8, wherein the first relevance is a value obtained by multiplying the first similarity by the recommendation level described for the itemized description portion corresponding to each of the first term groups. Further comprising a search unit,
At least one of the first term group and the second term group includes key terms related to serious situations;
the search unit searches for a term group including the key term from among the extracted first term group and the extracted second term group;
The medical support device according to claim 1 , wherein the output unit preferentially outputs, from the first literature information and the second literature information, an itemized description portion corresponding to the searched term group. The medical support device according to any one of claims 1 to 10, further comprising a behavior analysis unit that analyzes the behavior of the operator during output by the output unit, modifies the group of medical terms to add medical terms corresponding to the results of the analysis, and re-executes the first extraction unit, the second extraction unit, the third extraction unit, and the output unit in accordance with the modified group of medical terms. the plurality of literature information includes at least one guideline;
The itemized description portion in the guideline is a portion described by dividing it into the smallest items or clinical questions (CQs),
The medical support device according to any one of claims 1 to 11.