JP7278747B2 - Discharge summary generator - Google Patents

Description

An embodiment of the present invention relates to a discharge summary creation device.

A discharge summary is a summary of medical information for a patient's hospital stay. The hospital discharge summary is indispensable as a matter to be sent to a receiving institution, such as a rehabilitation facility, after discharge. A discharge summary is prepared by the attending physician.

It is known that creating a discharge summary is very burdensome for doctors. For example, when a doctor prepares a discharge summary, it is necessary to reread all the patient's hospital charts and extract important parts from the charts. In addition, when the patient's discharge from the hospital is suddenly decided, the doctor must promptly create a discharge summary. Under these circumstances, there is a demand for a discharge summary creation device that reduces the burden of creating a discharge summary.

JP 2014-071768 A JP 2010-224639 A

The problem to be solved by the present invention is to reduce the burden of creating a discharge summary.

A discharge summary creation device according to an embodiment includes an acquisition unit and a generation unit. The acquisition unit acquires a plurality of medical record data regarding a patient. The generation unit generates discharge summary data by inputting a plurality of medical chart data to a learning model that generates discharge summary data regarding a patient based on a plurality of medical chart data.

FIG. 1 is a diagram showing a configuration example of a medical information system according to the first embodiment. FIG. 2 is a diagram showing a configuration example of the discharge summary creating apparatus according to the first embodiment. FIG. 3 is a diagram for explaining a machine learning model and a learned model used in the discharge summary creation device according to the first embodiment. FIG. 4 is a diagram showing another configuration example of the discharge summary creating apparatus according to the first embodiment. FIG. 5 is a flowchart for explaining operations according to the first embodiment. FIG. 6 is a diagram for explaining typical contents of medical record data according to the first embodiment. FIG. 7 is a diagram for explaining the relationship between a plurality of medical record data and discharge summary data according to the first embodiment. FIG. 8 is a diagram showing a configuration example of a discharge summary creating apparatus according to the second embodiment. FIG. 9 is a flowchart for explaining operations according to the second embodiment. FIG. 10 is a diagram schematically showing the operation of the discharge summary creation device according to the second embodiment. FIG. 11 is a diagram showing a configuration example of a discharge summary creating apparatus according to the third embodiment. FIG. 12 is a flowchart for explaining operations according to the third embodiment. FIG. 13 is a diagram schematically showing the operation of the discharge summary creation device according to the third embodiment. FIG. 14 is a diagram showing a configuration example of a discharge summary creating apparatus according to the fourth embodiment. FIG. 15 is a flowchart for explaining operations according to the fourth embodiment. FIG. 16 is a diagram schematically showing the operation of the discharge summary creation device according to the fourth embodiment. FIG. 17 is a diagram showing a configuration example of a discharge summary creating apparatus according to the fifth embodiment. FIG. 18 is a flow chart for explaining the operation according to the fifth embodiment. FIG. 19 is a diagram for exemplifying a correspondence table in which appearance rates and words are associated with each other according to the fifth embodiment. FIG. 20 is a diagram for explaining the relationship between discharge summary data and updated discharge summary data according to the fifth embodiment. FIG. 21 is a diagram schematically showing the operation of the discharge summary creation device according to the fifth embodiment. FIG. 22 is a diagram schematically showing another operation of the discharge summary creation device according to the fifth embodiment. FIG. 23 is a diagram showing a configuration example of a medical information network according to the sixth embodiment. FIG. 24 is a flowchart for explaining operations according to the sixth embodiment. FIG. 25 is a diagram schematically showing the operation of the discharge summary creation device according to the sixth embodiment. FIG. 26 is a diagram for explaining a learning data set according to each embodiment.

Hereinafter, embodiments of the discharge summary creation device will be described in detail with reference to the drawings.

(First embodiment)
FIG. 1 is a diagram showing a configuration example of a medical information system according to the first embodiment. For example, as shown in FIG. 1 , a medical information system 10 according to the first embodiment includes a discharge summary creation device 11 , a hospital information system (HIS) 12 and a terminal device 13 . The discharge summary creation device 11, HIS 12, and terminal device 13 are communicably connected to each other via a network such as a LAN (Local Area Network). The medical information system 10 includes a radiology information system (RIS), a medical image diagnostic apparatus, a medical image management system (Picture Archiving and Communication Systems: PACS), and a data warehouse (DWH). etc. may be included.

The HIS 12 includes, for example, an electronic medical chart system that manages information on electronic medical charts. The electronic medical chart system includes an electronic medical chart DB (Data Base) 12a that stores information on electronic medical charts. Information related to electronic charts includes, for example, patient information and medical information. The HIS 12 may include a discharge summary creation device 11, which will be described later. Various functions of the discharge summary creating device 11, which will be described later, may be incorporated into the electronic medical record system.

Patient information is patient-specific information, and includes, for example, patient ID, patient name, gender, birthday and age. The medical information is information obtained by the medical staff regarding the patient's physical condition, disease state, treatment, etc. in the course of medical treatment. The medical information includes, for example, image information, examination history information, electrocardiogram information, vital sign information, medication history information, report information, nursing record information, and medical record information.

Image information is, for example, information representing the location of a medical image obtained by photographing a patient. The image information includes, for example, information representing the location of a medical image file generated by a medical image diagnostic apparatus as a result of an examination.

The test history information is, for example, information representing the history of test results acquired as a result of specimen tests, bacteriological tests, and the like performed on patients.

The electrocardiogram information is, for example, information about an electrocardiogram waveform measured from a patient.

Vital sign information is, for example, basic information related to a patient's life. Vital sign information includes, for example, pulse rate, respiration rate, oxygen concentration, body temperature, blood pressure and level of consciousness. The medication history information is, for example, information indicating the history of the amount of medication administered to the patient.

For example, the report information is obtained by reading medical images such as X-ray images, CT images, MRI images, and ultrasound images by a radiology interpreting doctor in response to an examination request from a clinician in a clinical department, and obtaining the patient's condition and It is information summarized about the disease. The report information includes, for example, interpretation report information representing an interpretation report created with reference to the medical image files stored in the PACS 5 by the interpretation doctor. Since the report information is generally stored in the PACS, the electronic medical record system can display the report information by reading the report information from the PACS.

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

The medical record description information is, for example, information input to the electronic medical chart by a medical doctor or the like. The medical record information includes, for example, medical records during hospitalization, patient's medical history and drug prescription history. Medical records during hospitalization are recorded, for example, in association with the input date and time. The medical record data in this embodiment is assumed to be, for example, the content described in medical records during hospitalization, but may include other information. Other information includes, for example, a patient's medical history and drug prescription history, medical images associated with medical records, or medical moving images. This also applies to each of the following embodiments.

The terminal device 13 corresponds to, for example, a PC (Personal Computer), a tablet PC, a PDA (Personal Digital Assistant), a smart phone, and the like. The terminal device 13 is arranged in each clinical department within the medical facility. Medical record description information is input to the terminal device 13 by a doctor.

In addition, the terminal device 13 can browse the electronic medical chart of any patient by accessing the electronic medical chart system installed in the HIS 12 . When creating a discharge summary, the doctor gives an instruction regarding the creation of the discharge summary from the viewing screen of the electronic medical record. The instruction regarding the preparation of the discharge summary is, for example, an instruction to start preparation of the discharge summary by the discharge summary preparation device 11, an instruction to designate a patient for which the discharge summary is prepared, and the like. The terminal device 13 transmits an instruction regarding creation of a discharge summary to the discharge summary creation device 11 .

FIG. 2 is a diagram showing a configuration example of the discharge summary creating apparatus according to the first embodiment. For example, as shown in FIG. 2, the discharge summary creation device 11 according to the first embodiment includes a processing circuit 21, a memory 22 and a communication interface 23. FIG. The processing circuit 21, the memory 22 and the communication interface 23 are communicably connected to each other via a bus, for example.

The memory 22 corresponds to, for example, a HDD (Hard Disk Drive), an SSD (Solid State Drive), an integrated circuit storage device, and the like. Also, the memory 22 may be a drive device or the like that reads and writes various information from/to a portable storage medium such as a CD-ROM drive, a DVD drive, and a flash memory.

The memory 22 stores, for example, a trained model 22a. The memory 22 may store the learned model 22a in advance when the discharge summary creation device 11 is shipped from the factory. Alternatively, the memory 22 may store the learned model 22a acquired from a server device (not shown) or the like after the discharge summary creation device 11 is shipped from the factory. This also applies to each of the following embodiments.

FIG. 3 is a diagram for explaining a machine learning model and a learned model used in the discharge summary creation device according to the first embodiment. For example, as shown in FIG. 3, a trained model 22a is a trained machine learning model obtained by causing a machine learning model 32 to perform machine learning according to a model learning program based on a learning data set 31. is. The learning data set 31 includes a plurality of medical record data as input data and discharge summary data as output data. The trained model 22a of this embodiment is configured to output discharge summary data 34 summarizing a plurality of medical chart data 33 based on the input of a plurality of medical chart data 33 .

More specifically, the machine learning model 32 is a parameterized composite function obtained by synthesizing a plurality of functions that receive a plurality of medical record data and output discharge summary data. A parameterized composite function is defined by a combination of multiple adjustable functions and parameters. The machine learning model according to the present embodiment may be any parameterized composite function that satisfies the above requirements, but shall be a multi-layered network model (hereinafter referred to as a multi-layered network).

The trained model 22a using a multi-layered network has an input layer that inputs a plurality of medical chart data 33, an output layer that outputs discharge summary data 34 corresponding to the plurality of medical chart data, and a layer between the input layer and the output layer. and at least one intermediate layer provided in the The trained model 22a is assumed to be used as a program module that is part of artificial intelligence software.

The communication interface 23 performs data communication with the HIS 12 and the terminal device 13 connected via the network. The communication standard between the HIS 12 and the terminal device 13 may be any standard, but examples thereof include standards conforming to HL7 (Health Level Seven) and DICOM (Digital Imaging and Communications in Medicine).

The processing circuit 21 has a processor (not shown), ROM (Read-Only Memory) and memory such as RAM, and controls the discharge summary creating apparatus 11 . The processing circuit 21 has an operation reception function 21a, a chart information acquisition function 21b, and a discharge summary generation function 21c. Various functions performed by the operation reception function 21a, the chart information acquisition function 21b, and the discharge summary generation function 21c are stored in the memory or the memory 22 in the form of programs executable by a computer. The processing circuit 21 is a processor that reads programs corresponding to these various functions from the memory and executes them to implement functions corresponding to each program. In other words, the processing circuit 21 in a state in which each program is read has a plurality of functions shown in the processing circuit 21 of FIG.

In the above description, it is assumed that the single processing circuit 21 implements the various functions described above, but the various functions may be implemented by a plurality of independent processors executing programs. In other words, the various functions described above may be configured as programs, and one processing circuit may execute each program, or a specific function may be implemented in a dedicated, independent program execution circuit. good too. Instead of storing the program in the memory 22 or the like, the program may be configured to be directly embedded in the circuit of the processor. In this case, the processor realizes its function by reading and executing the program embedded in the circuit.

The processor in the processing circuit 21 implements various functions by reading and executing a program stored in the memory 22 or the like, for example. Further, the operation reception function 21a, the chart information acquisition function 21b, and the hospital discharge summary generation function 21c of the processing circuit 21 are examples of the reception unit, the acquisition unit, and the generation unit, respectively.

The operation reception function 21 a is a function that receives an operator's (for example, a doctor's) operation via the terminal device 13 . For example, the processing circuit 21 receives an instruction regarding creation of a discharge summary from the terminal device 13 . Specifically, the processing circuit 21 receives, from the terminal device 13, an instruction to start creating a discharge summary, an instruction to designate a patient for whom the discharge summary is to be generated, and the like.

The chart information acquisition function 21b is a function for acquiring a plurality of chart data regarding a patient. Specifically, when accepting the designation of a patient, the processing circuit 21 acquires all hospitalized medical records of the patient stored in the HIS 12 as a plurality of chart data. That is, one piece of chart data corresponds to a medical record recorded by a doctor at a certain point in time for a certain patient.

The discharge summary generation function 21c is a function for outputting discharge summary data based on a plurality of acquired medical record data and a learned model. For example, the processing circuitry 21 generates discharge summary data by inputting multiple medical chart data to a learning model that generates discharge summary data for a patient based on multiple medical chart data.

FIG. 4 is a diagram showing another configuration example of the discharge summary creating apparatus according to the first embodiment. For example, as shown in FIG. 4, the processing circuitry 21 of the discharge summary creation device 11 may include a trained model 41 . That is, instead of storing the program (learned model 22a) in the memory 22, the program (learned model 41) may be directly embedded in the circuit of the processor. In this case, the processor implements the function by reading and executing the program (learned model 22a) incorporated in the circuit. This also applies to each of the following embodiments.

Next, the operation of the discharge summary creating apparatus 11 according to the first embodiment configured as described above will be described with reference to the flowchart of FIG.

FIG. 5 is a flowchart for explaining operations according to the first embodiment. In the process of the flowchart of FIG. 5, for example, when the discharge summary creation device 11 receives an instruction relating to creation of a discharge summary from the terminal device 13, the processing circuit 21 executes a program relating to creation of a discharge summary (discharge summary creation program). ) by executing

(Step S101)
When the discharge summary creation program is executed, the processing circuit 21 calls the operation reception function 21a. The processing circuit 21 receives input of patient information by the operation reception function 21a. The processing circuit 21 receives input of patient information output from the terminal device 13, for example.

(Step S102)
After receiving the patient information input, the processing circuit 21 calls the chart information acquisition function 21b. The processing circuit 21 acquires a plurality of medical chart data by the medical chart information acquisition function 21b. Specifically, the processing circuit 21 acquires a medical record corresponding to the received patient information, and acquires a plurality of chart data corresponding to each recorded date.

FIG. 6 is a diagram for explaining typical contents of medical record data according to the first embodiment. For example, as shown in FIG. 6 , medical chart data 61 includes patient information 62 , subjective information (Subjective) 63 , objective information (Objective) 64 , assessment (Assessment) 65 and plan (Plan) 66 . Typically, other than the patient information 62, it often takes the form of free description by the doctor. It shall be

The subjective information 63 describes, for example, the reason for consultation, chief complaint, history of present illness, and past history. Also, the subjective information 63 may describe family history such as close relatives, cohabiting family members, and key persons, and life history such as favorites, lifestyle habits, and social history. The objective information 64 describes, for example, physical findings and examination findings. In the evaluation 65, for example, a problem list listing assumed disease names (or diagnosed disease names) that are likely to pose a problem in examining a patient, and differential diagnoses and treatment policies for the assumed disease names listed in the problem list are described. . Plan 66 describes, for example, a diagnosis plan, a treatment plan, an explanation plan, and the like.

For example, the processing circuit 21 acquires chart data in which items of subjective information, objective information, evaluation, and planning are described.

(Step S103)
After obtaining a plurality of chart data, the processing circuit 21 calls the discharge summary generation function 21c. The processing circuit 21 uses the discharge summary generation function 21c to output discharge summary data based on the acquired plurality of medical record data and the learned model. Specifically, the processing circuit 21 reads out the learned model 22a from the memory 22, inputs a plurality of chart data to the learned model 22a, and outputs discharge summary data.

FIG. 7 is a diagram for explaining the relationship between a plurality of medical record data and discharge summary data according to the first embodiment. For example, as shown in FIG. 7, the plurality of medical chart data includes medical chart data 71, medical chart data 72 and medical chart data 73. FIG.

Part of the medical record data 71 includes, for example, "Broken left femur. Surgery was performed on MM/DD, bolts were inserted, and the fractured part was fixed. There were no symptoms of infection, and hospitalization was completed." The inside contains the sentence, "Instruct to fix the left leg and move in a wheelchair." A part of the medical record data 72 includes, for example, the sentence "During the follow-up, the pain seemed to be severe, so the patient was instructed to prescribe an analgesic. Afterwards, the patient recovered smoothly and the fracture was completely healed." ing. A part of the medical chart data 73 includes, for example, the sentence "The bolt extraction operation was performed on △ month △ day. It was completed without any problems, and the progress was good. No infection of the wound was induced." there is

The discharge summary creating apparatus 11 outputs discharge summary data 77 in response to inputs such as medical chart data 71, medical chart data 72, and medical chart data 73. FIG. The discharge summary data 77 includes an extracted portion 74 described in the chart data 71, an extracted portion 75 described in the chart data 72, and an extracted portion 76 described in the chart data 73, respectively. The content of the excerpt may be a summary of a part of the chart data.

The excerpt 74 includes, for example, the sentence "Fracture of the femur. A bolt was put in on a month and a day and an operation was performed." The excerpt 75 includes, for example, the sentence "Prescribe an analgesic. After that, the fracture is completely healed." The excerpt 76 includes, for example, the sentence "Bolt removal surgery was performed on △ month △ day. No infection occurred."

As described above, the discharge summary data is composed of sentences or sentences extracted from each of a plurality of medical chart data. The discharge summary data may be composed of sentences or sentences extracted from a part of medical chart data among a plurality of medical chart data. In addition, the discharge summary data may include at least one of medical images, medical moving images, and vital data in addition to character information.

As described above, according to the first embodiment, the discharge summary creation device 11 obtains a plurality of medical chart data regarding a patient, and uses a learning model for generating discharge summary data regarding the patient based on the plurality of medical chart data. and generate discharge summary data by entering multiple chart data.

Therefore, the discharge summary creation apparatus 11 according to the first embodiment automatically creates a discharge summary without setting extraction conditions, unlike the conventional method of weighting words and sentences in creating a discharge summary based on rules. can be created with As a result, the labor involved in creating and updating rules in the rule base can be reduced, and the burden on the administrator who manages the discharge summary creating apparatus 11 can be reduced. Therefore, the discharge summary creating apparatus 11 according to the first embodiment can reduce the burden of creating a discharge summary.

(Second embodiment)
In the first embodiment, the discharge summary creation device 11 creates discharge summary data from a plurality of medical record data. On the other hand, in the second embodiment, a case will be described in which discharge summary data is created in consideration of the destination of the discharge summary (for example, a referral hospital). In the following description, the same reference numerals are assigned to substantially the same parts as those in FIGS. 1 to 7, and the description thereof will be omitted, and mainly the different parts will be described.

FIG. 8 is a diagram showing a configuration example of a discharge summary creating apparatus according to the second embodiment. For example, as shown in FIG. 8, the discharge summary creation device 11 according to the second embodiment includes a processing circuit 21, a memory 22 and a communication interface 23. FIG. The processing circuit 21, the memory 22 and the communication interface 23 are communicably connected to each other via a bus, for example.

The processing circuit 21 has an operation reception function 21 a , a chart information acquisition function 21 b , a discharge summary generation function 21 c and a referral destination extraction function 81 . Various functions performed by the operation reception function 21a, the chart information acquisition function 21b, the discharge summary generation function 21c, and the referral destination extraction function 81 are stored in the memory in the form of programs executable by a computer. The referral destination extracting function 81 of the processing circuit is an example of an extractor.

The referral destination extraction function 81 is a function for extracting referral destinations from a plurality of chart data. Referral destinations correspond to other medical institutions. Other medical institutions include, for example, rehabilitation facilities, transfer destination hospitals, and the like. In addition, the referral destination may include inside the hospital. For example, when a certain patient leaves the hospital and switches to an outpatient consultation, it is assumed that the referral destination will be the hospital.

With the addition of the referral destination extracting function 81, the discharge summary generating function 21c is provided with the functions described below, or the functions described in the first embodiment are replaced with the functions described below.

The hospital discharge summary generation function 21c is a function for outputting hospital discharge summary data based on a plurality of acquired medical record data and a learned model corresponding to the extracted referral destination. For example, the processing circuit 21 generates discharge summary data by inputting a plurality of medical chart data to the learning model corresponding to the extracted referral destination.

A trained model 22a according to the second embodiment includes a plurality of trained models respectively corresponding to a plurality of referral destinations. Specifically, the learned model 22a includes, for example, a learned model for creating a discharge summary suitable for a rehabilitation facility, a learning model for creating a discharge summary suitable for a hospital to which a patient is transferred, and the like. include. A plurality of trained models are obtained by causing a machine learning model to perform machine learning according to a model learning program based on learning data sets prepared for each referral destination.

Next, the operation of the discharge summary creating apparatus 11 according to the second embodiment configured as described above will be described with reference to the flowchart of FIG.

FIG. 9 is a flowchart for explaining operations according to the second embodiment. The processing of the flowchart of FIG. 9 is started by the processing circuit 21 executing a discharge summary creation program, for example, when the discharge summary creation device 11 receives an instruction regarding creation of a discharge summary from the terminal device 13. be. Note that the processes of steps S201 and S202 are the same as the processes of steps S101 and S102 of FIG. 5, and thus description thereof is omitted.

(Step S203)
After obtaining a plurality of chart data, the processing circuit 21 calls the referral destination extraction function 81 . The processing circuit 21 uses a referral destination extraction function 81 to extract referral destinations from a plurality of chart data.

(Step S203)
After extracting referral destinations from a plurality of chart data, the processing circuit 21 calls the discharge summary generation function 21c. The processing circuit 21 uses the discharge summary generation function 21c to output discharge summary data based on the acquired plurality of medical record data and the learned model corresponding to the extracted referral destination. Specifically, the processing circuit 21 reads the learned model 22a corresponding to the referral destination from the memory 22, inputs a plurality of medical record data to the learned model 22a, and outputs discharge summary data. .

FIG. 10 is a diagram schematically showing the operation of the discharge summary creation device according to the second embodiment. For example, as shown in FIG. 10, the discharge summary creation device 11 according to the second embodiment includes a first trained model 102, a second trained model 103, and a third trained model as a plurality of trained models. It has a finished model 104 .

When a plurality of medical chart data 101 are input, the discharge summary creating apparatus 11 extracts referral destinations associated with the first trained model 102 from the plurality of medical chart data 101, for example. Then, the hospital discharge summary creation device 11 creates discharge summary data 105 based on the plurality of chart data 101 and the first learned model 102 . Therefore, the discharge summary creation device 11 can create the discharge summary data 105 for the referral destination associated with the first trained model 102 .

As described above, according to the second embodiment, the discharge summary creation device 11 has a plurality of learning models corresponding to a plurality of referral destinations, acquires a plurality of medical chart data about the patient, and obtains a plurality of medical chart data. A referral destination is extracted from the hospital, and discharge summary data is generated by inputting a plurality of chart data to a learning model corresponding to the extracted referral destination.

Therefore, the discharge summary creating apparatus 11 according to the second embodiment can automatically create a discharge summary that matches the referral destination of the discharged patient.

(Third embodiment)
In the second embodiment, a case has been described in which the discharge summary creation device 11 extracts referral destinations and creates a discharge summary suitable for the extracted referral destinations. On the other hand, in the third embodiment, a case will be described in which a doctor or the like receives information on a referral destination and creates a discharge summary suitable for the referral destination. 1 to 10 are denoted by the same reference numerals, and description thereof will be omitted, and different portions will be mainly described.

FIG. 11 is a diagram showing a configuration example of a discharge summary creating apparatus according to the third embodiment. For example, as shown in FIG. 11, a discharge summary creation device 11 according to the third embodiment includes a processing circuit 21, a memory 22 and a communication interface 23. FIG. The processing circuit 21, the memory 22 and the communication interface 23 are communicably connected to each other via a bus, for example.

The processing circuit 21 has an operation reception function 21 a , a chart information acquisition function 21 b , a discharge summary generation function 21 c and a referral designation function 111 . Various functions performed by the operation reception function 21a, the chart information acquisition function 21b, the discharge summary generation function 21c, and the referral destination designation function 111 are stored in the memory in the form of programs executable by a computer. The referral destination specifying function 111 included in the processing circuit is an example of a specifying unit.

In the operation reception function 21 a according to the third embodiment, the processing circuit 21 receives input of introduction destination information (introduction destination information) from the terminal device 13 .

The referral destination designation function 111 is a function for designating a referral destination. Specifically, the processing circuit 21 specifies to use the learned model corresponding to the received referral destination from among the plurality of learned models.

Due to the addition of the referral destination designation function 111, the discharge summary generation function 21c is provided with the functions described below, or the functions described in the first embodiment are replaced with the functions described below.

The hospital discharge summary generation function 21c is a function for outputting hospital discharge summary data based on a plurality of acquired medical record data and a trained model corresponding to a designated referral destination. For example, the processing circuit 21 generates hospital discharge summary data by inputting a plurality of medical record data to a learning model corresponding to a referral destination.

A trained model 22a according to the third embodiment includes a plurality of trained models respectively corresponding to a plurality of referral destinations. A plurality of trained models are obtained by causing a machine learning model to perform machine learning according to a model learning program based on learning data sets prepared for each referral destination.

Next, the operation of the discharge summary creation apparatus 11 according to the third embodiment configured as described above will be described with reference to the flowchart of FIG.

FIG. 12 is a flowchart for explaining operations according to the third embodiment. The processing of the flowchart of FIG. 12 is started by the processing circuit 21 executing the discharge summary creation program when the discharge summary creation device 11 receives an instruction regarding the creation of the discharge summary from the terminal device 13 . Note that the processes of steps S301 and S302 are the same as the processes of steps S101 and S102 of FIG. 5, and therefore description thereof is omitted.

(Step S303)
After acquiring a plurality of chart data, the processing circuit 21 calls the referral destination specifying function 111 . The processing circuit 21 receives input of referral destination information by the operation reception function 21a. Then, the processing circuit 21 uses the referral destination designating function 111 to designate a trained model corresponding to the received referral destination from among a plurality of trained models to be used.

(Step S304)
After designating the referral destination, the processing circuit 21 calls the discharge summary generation function 21c. The processing circuit 21 uses the discharge summary generation function 21c to output discharge summary data based on a plurality of acquired medical chart data and a trained model corresponding to the designated referral destination. Specifically, the processing circuit 21 reads out the learned model 22a corresponding to the referral destination from the memory 22, and generates discharge summary data by inputting a plurality of chart data to the learned model.

FIG. 13 is a diagram schematically showing the operation of the discharge summary creation device according to the third embodiment. For example, as shown in FIG. 13, the discharge summary creation device 11 according to the third embodiment includes a first trained model 102, a second trained model 103, and a third trained model as a plurality of trained models. It has a finished model 104 .

When a plurality of medical record data 101 and referral destination information 131 are input, the discharge summary creating apparatus 11 designates, for example, the first trained model 102 corresponding to the referral destination included in the referral destination information 131 . Then, the hospital discharge summary creation device 11 creates discharge summary data 105 based on the plurality of chart data 101 and the first learned model 102 . Therefore, the discharge summary creation device 11 can create the discharge summary data 105 for the referral destination associated with the first trained model 102 .

As described above, according to the third embodiment, the discharge summary creation device 11 has a plurality of learning models corresponding to a plurality of referral destinations, acquires a plurality of medical record data regarding the patient, and obtains a plurality of patient referral destinations. is set, and discharge summary data is generated by inputting a plurality of medical record data to the learning model corresponding to the set referral destination.

Therefore, the discharge summary creating apparatus 11 according to the third embodiment can create a discharge summary that matches the referral destination of the discharged patient.

(Fourth embodiment)
In the third embodiment, a case has been described where information on a referral destination is received and a discharge summary suitable for the referral destination is created. On the other hand, in the fourth embodiment, a case will be described in which a plurality of discharge summaries are created and an arbitrary discharge summary is selected from among the created plurality of discharge summaries. In the following description, substantially the same parts as those in FIGS. 1 to 13 are denoted by the same reference numerals, and descriptions thereof are omitted, and different parts are mainly described.

FIG. 14 is a diagram showing a configuration example of a discharge summary creating apparatus according to the fourth embodiment. For example, as shown in FIG. 14, the discharge summary creation device 11 according to the fourth embodiment includes a processing circuit 21, a memory 22 and a communication interface 23. FIG. The processing circuit 21, the memory 22 and the communication interface 23 are communicably connected to each other via a bus, for example.

The processing circuit 21 has an operation reception function 21 a , a chart information acquisition function 21 b , a discharge summary generation function 21 c and a discharge summary selection function 141 . Various functions performed by the operation reception function 21a, the chart information acquisition function 21b, the discharge summary generation function 21c, and the discharge summary selection function 141 are stored in the memory in the form of programs executable by a computer. The discharge summary selection function 141 included in the processing circuit is an example of a selection unit.

In the operation reception function 21a according to the fourth embodiment, the processing circuit 21 receives an input regarding selection of a discharge summary from the terminal device 13. FIG.

The discharge summary selection function 141 is a function for selecting an arbitrary discharge summary from a plurality of discharge summaries. Specifically, the processing circuitry 21 selects an arbitrary discharge summary from a plurality of discharge summaries according to instructions regarding the selection of the discharge summary input by the doctor.

With the addition of the discharge summary selection function 141, in the discharge summary generation function 21c, the functions described below are added, or the functions described in the first embodiment are replaced with the functions described below.

The discharge summary generation function 21c is a function for outputting a plurality of discharge summary data based on a plurality of obtained medical record data and a plurality of trained models. For example, the processing circuit 21 outputs a plurality of discharge summary data by inputting a plurality of medical record data to a plurality of trained models. It contains multiple pre-trained models corresponding to each other. A plurality of trained models are obtained by causing a machine learning model to perform machine learning according to a model learning program based on learning data sets prepared for each referral destination.

Next, the operation of the discharge summary creating apparatus 11 according to the fourth embodiment configured as described above will be described with reference to the flowchart of FIG.

FIG. 15 is a flowchart for explaining operations according to the fourth embodiment. The processing of the flowchart of FIG. 15 is started by the processing circuit 21 executing the discharge summary creation program when the discharge summary creation device 11 receives an instruction regarding the creation of the discharge summary from the terminal device 13 . Note that the processes of steps S401 and S402 are the same as the processes of steps S401 and S402 of FIG. 5, and thus description thereof is omitted.

(Step S403)
After obtaining a plurality of chart data, the processing circuit 21 calls the discharge summary generation function 21c. The processing circuit 21 uses the discharge summary generation function 21c to output a plurality of discharge summary data based on the acquired plurality of medical record data and the plurality of learned models. Specifically, the processing circuit 21 reads out a plurality of learned models from the memory 22 and inputs a plurality of chart data to the learned models to generate a plurality of hospital discharge summary data.

(Step S404)
After generating the plurality of discharge summary data, processing circuitry 21 calls discharge summary selection function 141 . The processing circuit 21 receives a selection of desired discharge summary data from a plurality of discharge summary data, and outputs the selected discharge summary data.

Specifically, the processing circuit 21 transmits a plurality of discharge summary data to the terminal device 13 operated by the doctor. The doctor selects arbitrary discharge summary data from a plurality of discharge summary data displayed on the terminal device 13 . The terminal device 13 transmits to the processing circuit 21 instructions regarding the selection of the discharge summary entered by the doctor. The processing circuit 21 receives an instruction regarding selection of a discharge summary, and outputs discharge summary data selected according to the instruction.

FIG. 16 is a diagram schematically showing the operation of the discharge summary creation device according to the fourth embodiment. For example, as shown in FIG. 16, the discharge summary creation device 11 according to the fourth embodiment includes a first trained model 102, a second trained model 103, and a third trained model as a plurality of trained models. It has a finished model 104 .

When a plurality of medical chart data 101 are input, the discharge summary creating apparatus 11 creates first discharge summary data 161 based on the plurality of medical chart data 101 and the first trained model 102 . Similarly, the discharge summary creation device 11 creates second discharge summary data 162 based on the plurality of medical chart data 101 and the second learned model, and creates the plurality of medical chart data 101 and the third learned model. Based on the model 104, a third discharge summary data 163 is created.

After creating first discharge summary data 161 , second discharge summary data 162 and third discharge summary data 163 , discharge summary creation device 11 outputs these discharge summary data to terminal device 13 . The discharge summary creation device 11 receives user selection information 164 which is an instruction from a doctor (user) operating the terminal device 13 . The user selection information 164 is, for example, information instructing selection of the first discharge summary data 161 from among the first discharge summary data 161, the second discharge summary data 162, and the third discharge summary data 163. be. The discharge summary creation device 11 outputs the first discharge summary data 161 as the discharge summary data 105 according to the user selection information 164 .

As described above, according to the fourth embodiment, the discharge summary creation device 11 is provided with a plurality of learning models corresponding to a plurality of referral destinations, acquires a plurality of medical chart data related to the patient, and acquires a plurality of learning models. , generate multiple discharge summary data by inputting multiple chart data.

Therefore, the discharge summary creating apparatus 11 according to the fourth embodiment can create a discharge summary that matches the referral destination of the discharged patient.

(Fifth embodiment)
In the first to fourth embodiments, the discharge summary creation device 11 creates discharge summary data from a plurality of medical record data. On the other hand, in the fifth embodiment, a case will be described in which a more complete discharge summary is created by updating the document using a plurality of discharge summary data that match the referral destination. 1 to 16 are denoted by the same reference numerals, and description thereof will be omitted, and different portions will be mainly described.

FIG. 17 is a diagram showing a configuration example of a discharge summary creating apparatus according to the fifth embodiment. For example, as shown in FIG. 17, the discharge summary creation device 11 according to the fifth embodiment includes a processing circuit 21, a memory 22 and a communication interface 23. FIG. The processing circuit 21, the memory 22 and the communication interface 23 are communicably connected to each other via a bus, for example.

The processing circuit 21 has an operation reception function 21 a , a chart information acquisition function 21 b , a discharge summary generation function 21 c , an importance calculation function 171 , a referral destination word extraction function 172 and a discharge summary update function 173 . Various functions performed by the operation reception function 21a, the medical record information acquisition function 21b, the discharge summary generation function 21c, the importance calculation function 171, the referral target word extraction function 172, and the discharge summary update function 173 are implemented by a computer-executable program. stored in memory in the form The importance calculation function 171, the referral destination word extraction function 172, and the discharge summary update function 173 of the processing circuit are examples of a calculation unit, a word extraction unit, and an update unit.

The importance calculation function 171 is a function for calculating the importance of each word included in a plurality of chart data. Specifically, the processing circuit 21 extracts a plurality of words from a plurality of chart data, and calculates the importance of the plurality of words based on the appearance rate (appearance frequency) of the plurality of words.

As the degree of importance, for example, TF (Term Frequency) based on the appearance frequency of a word in a document may be used, or IDF (Inverse Document Frequency), which is the reciprocal of the document frequency in which a word appears, may be used. or both (TF-IDF) may be used.

The referral destination word extraction function 172 is a function for extracting a plurality of frequently appearing words related to the referral destination from a plurality of discharge summary data suitable for the referral destination. The plurality of frequently appearing words correspond to, for example, the words with the highest number of appearances (for example, up to 10th place) among the extracted words. For example, if the referral destination is a rehabilitation facility, the hospital discharge summary data suitable for the referral destination is the hospital discharge summary data for the rehabilitation facility.

The discharge summary update function 173 is a function for updating the discharge summary data based on the discharge summary data, multiple medical chart data, importance, and multiple frequent words. The hospital discharge summary data is the same as that output in the first to fourth embodiments described above. In this embodiment, updated discharge summary data is output by updating the discharge summary data. Updating corresponds to, for example, increasing the contents described in the discharge summary or adding sentences. At least one of medical images, medical moving images, and vital data may be included as the update contents.

Next, the operation of the discharge summary creating apparatus 11 according to the fifth embodiment configured as described above will be described with reference to the flowchart of FIG.

FIG. 18 is a flow chart for explaining the operation according to the fifth embodiment. The processing of the flowchart of FIG. 18 is started by the processing circuit executing a discharge summary creation program, for example, when the discharge summary creation device 11 receives an instruction regarding creation of a discharge summary from the terminal device 13. . Since steps S501 to S503 are the same as steps S101 to S103 in FIG. 5, the description thereof is omitted.

(Step S504)
After outputting the discharge summary data, the processing circuit 21 calls the importance calculation function 171 . The processing circuit 21 uses the degree-of-importance calculation function 171 to extract a plurality of words from a plurality of chart data.

(Step S505)
After extracting the plurality of words in step S504, processing circuit 21 calculates the importance of each of the plurality of words based on the appearance rate of the plurality of words. Combinations of appearance rates and words are stored in, for example, the memory 22 as a correspondence table.

FIG. 19 is a diagram for exemplifying a correspondence table in which appearance rates and words are associated with each other according to the fifth embodiment. For example, as shown in FIG. 19, the correspondence table 190 associates the appearance rate with the word. The appearance rate is, for example, the ratio of medical chart data containing a certain word to a plurality of medical chart data. For example, the appearance rate of the word “fracture” of “90%” means that 90% of the plurality of medical chart data contain the word “fracture”. In the description of this flowchart, importance is synonymous with appearance rate.

(Step S506)
After calculating the importance of each of the plurality of words, the processing circuit 21 calls the introduction destination word extraction function 172 . The processing circuit 21 uses the referral destination word extraction function 172 to extract a plurality of frequently appearing words related to the referral destination from a plurality of discharge summary data suitable for the referral destination.

(Step S507)
After extracting a plurality of frequent words, processing circuit 21 calls update discharge summary function 173 . The processing circuit 21 uses the discharge summary update function 173 to update the discharge summary data based on the discharge summary data, the plurality of medical record data, the degree of importance, and the plurality of frequently appearing words.

Specifically, the processing circuit 21 uses the discharge summary update function 173 to extract a plurality of frequently appearing words whose importance exceeds the threshold (hereinafter referred to as extracted words). The processing circuit 21 extracts sentences containing extracted words from a plurality of medical record data, and adds the extracted sentences to the discharge summary data, thereby updating the discharge summary data. In addition, when the extracted sentence overlaps with the discharge summary data, the processing circuit 21 may not add the overlapping sentence to the discharge summary data. Also, the extracted sentence may be replaced with a sentence of high importance.

FIG. 20 is a diagram for explaining the relationship between discharge summary data and updated discharge summary data according to the fifth embodiment. For example, as shown in FIG. 20, updated discharge summary data 202 is generated from discharge summary data 201 by updating the discharge summary. The updated hospital discharge summary data 202 corresponds to the hospital discharge summary data 201 to which an additional sentence 203 with a high degree of importance is added.

FIG. 21 is a diagram schematically showing the operation of the discharge summary creation device according to the fifth embodiment. For example, as shown in FIG. 21, the discharge summary creation device 11 according to the fifth embodiment includes a learned model 211, an importance calculation function 213, a referral target word extraction function 215, and a discharge summary update function 216. The learned model 211, the importance calculation function 213, the referral target word extraction function 215, and the discharge summary update function 216 are based on the learned model 22a, the importance calculation function 171, the referral target word extraction function 172, and the discharge summary of FIG. Since it is the same as the update function 173, detailed description is omitted. Also, the plurality of discharge summary data 214 is a plurality of discharge summary data suitable for the referral destination.

When a plurality of medical chart data 101 are input, the hospital discharge summary creation device 11 creates discharge summary data 212 based on the plurality of medical chart data 101 and the learned model 211 . The hospital discharge summary creation device 11 calculates the importance of each of the plurality of words from the plurality of medical chart data 101 by the importance calculation function 213 . When a plurality of discharge summary data 214 are further input, the discharge summary creation device 11 extracts a plurality of frequently appearing words from the plurality of discharge summary data 214 by a referral destination word extraction function 215 . The discharge summary creation device 11 uses the discharge summary update function 216 to update the discharge summary data 212 based on the discharge summary data 212, the plurality of medical record data 101, the degree of importance, and the plurality of frequently occurring words. The discharge summary creation device 11 outputs the updated discharge summary data 212 as updated discharge summary data 217 .

As described above, according to the fifth embodiment, the discharge summary creation device 11 obtains a plurality of patient-related chart data, and uses a learning model to generate discharge summary data related to the patient based on the plurality of chart data. and generate discharge summary data by entering multiple chart data. Then, the discharge summary creation device 11 calculates the degree of importance of each of the plurality of words included in the plurality of medical record data, and extracts a plurality of frequently appearing words related to the referral destination from the plurality of discharge summary data suitable for the referral destination. , based on the generated discharge summary data, multiple chart data, importance and multiple frequent words, update the discharge summary data.

Therefore, the hospital discharge summary creating apparatus 11 according to the fifth embodiment can add sentences containing words of high importance to the hospital discharge summary, thereby creating a more complete hospital discharge summary. be able to.

(Other operation examples)
FIG. 22 is a diagram schematically showing another operation of the discharge summary creation device according to the fifth embodiment. In FIG. 21, the case where the updated discharge summary data 217 is output by updating the discharge summary data 212 output from the learned model 211 has been described. On the other hand, FIG. 22 describes a case where updated discharge summary data 223 is output by updating previously generated discharge summary data 221, which is discharge summary data generated in advance. The pre-created discharge summary data 221 corresponds to discharge summary data generated by a doctor, for example.

The discharge summary creation device 11 uses the discharge summary update function 222 to update the pre-created discharge summary data 221 based on the pre-created discharge summary data 221, multiple medical chart data 101, importance and multiple frequently occurring words. The discharge summary creation device 11 outputs the updated pre-created discharge summary data 221 as updated discharge summary data 223 .

That is, the discharge summary creating apparatus 11 can add sentences of high importance to externally input data (pre-created discharge summary data 221).

(Sixth embodiment)
In the fifth embodiment, a case has been described in which a plurality of discharge summaries corresponding to referral destinations are used to create a discharge summary suitable for a referral destination. On the other hand, in the sixth embodiment, a case will be described in which a discharge summary suitable for a referral destination is created by using the referral destination's electronic medical record. The configuration of the discharge summary creation device 11 according to the sixth embodiment is the same as the configuration of the discharge summary creation device 11 according to the fifth embodiment shown in FIG. 1 to 22 are denoted by the same reference numerals, and descriptions thereof are omitted, and different portions are mainly described.

FIG. 23 is a diagram showing a configuration example of a medical information network according to the sixth embodiment. For example, as shown in FIG. 23, medical information network 230 includes medical information system 10 , medical information system 231 and medical information system 232 . These multiple medical information systems correspond to different hospital systems, for example. For example, when the discharge summary creation device 11 included in the medical information system 10 acquires data (eg, electronic medical record data) of another hospital (eg, the hospital building the medical information system 231), the discharge summary creation device 11 can obtain electronic medical record data from the HIS in the medical information system 231 via the medical information network 230 .

In the introduction destination word extraction function 172 according to the sixth embodiment, the processing circuit 21 extracts a plurality of frequently appearing words from a plurality of chart data created by the introduction destination.

FIG. 24 is a flowchart for explaining operations according to the sixth embodiment. The processing of the flowchart of FIG. 24 is started by the processing circuit executing a discharge summary creation program, for example, when the discharge summary creation device 11 receives an instruction regarding creation of a discharge summary from the terminal device 13. . Since steps S601 to S605 and step S607 are the same as steps S501 to S505 and step S507 in FIG. 18, description thereof is omitted. The learned model 211, the importance calculation function 213, the referral target word extraction function 215, and the discharge summary update function 216 are based on the learned model 22a, the importance calculation function 171, the referral target word extraction function 172, and the discharge summary of FIG. Since it is the same as the update function 173, detailed description is omitted. Also, the plurality of discharge summary data 214 is a plurality of discharge summary data suitable for the referral destination.

(Step S606)
After calculating the importance of each of the plurality of words, the processing circuit 21 calls the introduction destination word extraction function 172 . The processing circuit 21 uses the referral destination word extraction function 172 to extract a plurality of frequently appearing words related to the referral destination from a plurality of chart data created by the referral destination.

FIG. 25 is a diagram schematically showing the operation of the discharge summary creation device according to the sixth embodiment. For example, as shown in FIG. 25, the discharge summary creation device 11 according to the sixth embodiment includes a learned model 211, an importance calculation function 213, a referral target word extraction function 215, and a discharge summary update function 216. The learned model 211, the importance calculation function 213, the referral target word extraction function 215, and the discharge summary update function 216 are based on the learned model 22a, the importance calculation function 171, the referral target word extraction function 172, and the discharge summary of FIG. Since it is the same as the update function 173, detailed description is omitted. The referral destination electronic medical record data 251 is a plurality of electronic medical chart data created at the referral destination hospital (or rehabilitation facility, etc.).

When a plurality of medical chart data 101 are input, the hospital discharge summary creation device 11 creates discharge summary data 212 based on the plurality of medical chart data 101 and the learned model 211 . The hospital discharge summary creation device 11 calculates the importance of each of the plurality of words from the plurality of medical chart data 101 by the importance calculation function 213 . When the referral destination electronic medical record data 251 is further input, the discharge summary creating apparatus 11 extracts a plurality of frequently appearing words from the referral destination electronic medical chart data 251 by the referral destination word extraction function 215 . The discharge summary creation device 11 uses the discharge summary update function 216 to update the discharge summary data 212 based on the discharge summary data 212, the plurality of medical record data 101, the degree of importance, and the plurality of frequently occurring words. The discharge summary creation device 11 outputs the updated discharge summary data 212 as updated discharge summary data 252 .

As described above, according to the sixth embodiment, the discharge summary creation device 11 acquires a plurality of patient-related chart data, and uses the learning model to generate the patient-related discharge summary data based on the plurality of chart data. and generate discharge summary data by entering multiple chart data. Then, the discharge summary creation device 11 calculates the degree of importance of each of the plurality of words included in the plurality of medical record data, and extracts a plurality of frequently occurring words related to the referral destination from the plurality of medical chart data created by the referral destination. , based on the generated discharge summary data, multiple chart data, importance and multiple frequent words, update the discharge summary data.

Therefore, the hospital discharge summary creating apparatus 11 according to the sixth embodiment can add sentences containing words of high importance to the hospital discharge summary, thereby creating a more complete hospital discharge summary. be able to.

(Application example)
The discharge summary creation device 11 according to the first to sixth embodiments can create discharge summary data (or updated discharge summary data) from a plurality of medical record data. That is, each time the discharge summary creation device 11 is used, a learning data set, which is a combination of a plurality of medical record data and the created discharge summary data, can be obtained.

FIG. 26 is a diagram for explaining a learning data set according to each embodiment. For example, as shown in FIG. 26, when a plurality of medical record data 261 are input, the discharge summary creation device 11 combines the plurality of medical record data 261 with the learned model stored in the discharge summary creation device 11. Based on this, discharge summary data 262 is created. A plurality of medical record data 261 and hospital discharge summary data 262 can be used, for example, as a learning data set 263 for additional learning of a trained model. When the created hospital discharge summary data is used as a learning data set, a doctor may examine the content of the hospital discharge summary data and make corrections, additions, and the like.

According to at least one embodiment described above, it is possible to reduce the burden of creating a discharge summary.

The term "processor" in each of the above embodiments includes, for example, a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), an application specific integrated circuit (ASIC), a programmable logic device (e.g., Circuits such as Simple Programmable Logic Devices (SPLDs), Complex Programmable Logic Devices (CPLDs), and Field Programmable Gate Arrays (FPGAs).

While several embodiments of the invention have been described, these embodiments have been presented by way of example and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, replacements, and modifications can be made without departing from the scope of the invention. These embodiments and their modifications are included in the scope and spirit of the invention, as well as the scope of the invention described in the claims and equivalents thereof.

10 Medical information system 61, 71, 72, 73 Medical chart data 74, 75, 76 Excerpts 77, 201 Discharge summary data 190 Correspondence table 202 Updated discharge summary data 203 Additional text 230 Medical information network

Claims (5)

an acquisition unit that acquires a plurality of patient chart data;
a generation unit that generates the discharge summary data by inputting the plurality of medical record data to a learning model that generates discharge summary data regarding the patient based on the plurality of medical record data;
Multiple learning models corresponding to multiple referral destinations,
an extraction unit that extracts the referral destination from the plurality of medical record data;
and
The discharge summary data is created by summarizing the description contents of at least a part of the plurality of medical chart data ,
The generation unit generates the discharge summary data by inputting the plurality of medical record data into the learning model corresponding to the extracted referral destination.
Discharge summary generator. an acquisition unit that acquires a plurality of patient chart data;
a generation unit that generates the discharge summary data by inputting the plurality of medical chart data to a learning model that generates discharge summary data regarding the patient based on the plurality of medical chart data;
a calculation unit that calculates the importance of each of the plurality of words included in the plurality of medical record data;
a word extraction unit that extracts a plurality of frequently occurring words related to the referral destination from a plurality of discharge summary data that match the referral destination;
an updating unit that updates the discharge summary data based on the discharge summary data, the plurality of medical chart data, the importance level, and the plurality of frequent words;
and
The hospital discharge summary creation device, wherein the discharge summary data is created by summarizing description contents of at least a part of the plurality of medical chart data. an acquisition unit that acquires a plurality of patient chart data;
a generation unit that generates the discharge summary data by inputting the plurality of medical chart data to a learning model that generates discharge summary data regarding the patient based on the plurality of medical chart data;
a calculation unit that calculates the importance of each of the plurality of words included in the plurality of medical record data;
a word extraction unit that extracts a plurality of frequently occurring words related to the referral destination from a plurality of chart data of the referral destination;
an updating unit that updates the discharge summary data based on the discharge summary data, the plurality of medical chart data, the importance level, and the plurality of frequent words;
and
The hospital discharge summary creation device, wherein the discharge summary data is created by summarizing description contents of at least a part of the plurality of medical chart data. The update unit extracts words whose importance exceeds a threshold from the plurality of frequently appearing words, extracts sentences containing the words extracted from the plurality of medical record data, and extracts from the discharge summary data. 4. The discharge summary creation device according to claim 2 or 3 , wherein the discharge summary data is updated by adding the sentence written by the user. 5. The discharge summary creating apparatus according to claim 1 , wherein said discharge summary data includes at least one of medical images, medical moving images and vital data.

Images