Medical data processing system, medical data processing device, medical data processing method, program, and recording medium

Abstract

To improve the quality of medical data.SOLUTION: A medical data processing system includes an evaluation section. The evaluation section is configured to evaluate consistency between first medical data and second medical data. The first medical data contains any of first input data input by a medical worker, first examination data acquired on the basis of data acquired by a medicine examination, and previously prepared standard data. The second medical data contains any of second input data input by the medical worker and second examination data acquired on the basis of the data acquired by the medicine examination.SELECTED DRAWING: Figure 2

Description

Technical Field

[0001] The present disclosure relates to a medical data processing system, a medical data processing apparatus, a medical data processing method, a program, and a recording medium.

Background Art

[0002] In the medical field, a wide variety of data is generated and referenced. Just to mention a few examples, there are electronic medical records, surgical plans, surgical records, treatment plans, inpatient treatment plans, prescriptions, inspection orders, medical accounting data, nursing records, types (model numbers) of artificial objects transplanted into patients, measurement data of patients, photographed images, biochemical test data, biopsy data, genetic test data, clinical practice guidelines for specific diseases, clinical pathways, normative data, and the like.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] One object of the present disclosure is to improve the quality of medical data.

Means for Solving the Problems

[0005] One embodiment is a system for processing medical data, including an evaluation unit that executes an evaluation of the consistency between first medical data and second medical data, wherein the first medical data includes any one of first input data input by medical staff, first inspection data obtained based on data obtained by a medical examination, and pre-created standard data, and the second medical data includes any one of second input data input by medical staff and second inspection data obtained based on data obtained by a medical examination.

[0006] Another embodiment is a device for processing medical data, comprising a receiving unit for receiving first medical data and second medical data, and an evaluation unit for performing an evaluation of the consistency between the first medical data and the second medical data, wherein the first medical data includes any of the following: first input data entered by a medical professional, first test data obtained based on data obtained by a medical examination, and pre-created standard data, and the second medical data includes any of the following: second input data entered by a medical professional and second test data obtained based on data obtained by a medical examination.

[0007] Another embodiment is a method for processing medical data to support patient care, comprising: preparing first medical data including either first input data entered by a healthcare professional, first test data obtained based on data obtained by medical tests, and pre-created standard data; preparing second medical data including either second input data entered by a healthcare professional and second test data obtained based on data obtained by medical tests; and performing an evaluation of consistency between the first medical data and the second medical data.

[0008] Another embodiment is a program that causes a computer to process medical data to support the treatment of a patient, the program causing the computer to perform the following steps: prepare first medical data including first input data entered by a medical professional, first test data obtained based on data obtained by a medical examination, and pre-created standard data; prepare second medical data including second input data entered by a medical professional and second test data obtained based on data obtained by a medical examination; and perform an evaluation of the consistency between the first medical data and the second medical data.

[0009] Another embodiment is a computer-readable non-temporary recording medium on which a program is recorded that causes a computer to process medical data to support the treatment of a patient, the program causing the computer to perform the following steps: prepare first medical data including any of first input data entered by a medical professional, first test data obtained based on data obtained by a medical examination, and pre-created standard data; prepare second medical data including any of second input data entered by a medical professional and second test data obtained based on data obtained by a medical examination; and perform an evaluation of the consistency between the first medical data and the second medical data. [Effects of the Invention]

[0010] According to this embodiment, it is possible to improve the quality of medical data. [Brief explanation of the drawing]

[0011] [Figure 1] This is a block diagram illustrating an example of an in-house information system for a medical institution, including a medical data processing system according to an exemplary embodiment of the model. [Figure 2] This is a block diagram showing the configuration of a medical data processing system according to an exemplary embodiment of the system. [Figure 3] This is a table showing the types of medical data processed by a medical data processing system according to an exemplary embodiment of the model. [Figure 4] This is a block diagram showing the configuration of a medical data processing system according to an exemplary embodiment of the system. [Figure 5] This is a block diagram illustrating the configuration of a medical data processing system according to an exemplary embodiment of the system. [Figure 6] This is a block diagram showing the configuration of a medical data processing system according to an exemplary embodiment of the system. [Figure 7] This is a flowchart illustrating the operation of a medical data processing system according to an exemplary embodiment of the system. [Figure 8] A flowchart for explaining the operation of a medical data processing system according to an exemplary aspect of an embodiment. [Figure 9] A flowchart for explaining the operation of a medical data processing system according to an exemplary aspect of an embodiment. [Figure 10] A flowchart for explaining the operation of a medical data processing system according to an exemplary aspect of an embodiment. [Figure 11] A flowchart for explaining the operation of a medical data processing system according to an exemplary aspect of an embodiment. [Figure 12] A flowchart for explaining the operation of a medical data processing system according to an exemplary aspect of an embodiment. [Figure 13] A flowchart for explaining the operation of a medical data processing system according to an exemplary aspect of an embodiment. [Figure 14] A flowchart for explaining the operation of a medical data processing system according to an exemplary aspect of an embodiment.

MODE FOR CARRYING OUT THE INVENTION

[0012] Non-limiting embodiments according to the present disclosure will be described in detail with reference to the drawings.

[0013] In some exemplary aspects of the embodiments according to the present disclosure, applications to the field of ophthalmology will be described. However, those skilled in the art will understand that the application fields of the embodiments are not limited to ophthalmology, and applications to the general medical field (for example, any medical department) are possible.

[0014] Any known technology can be combined with the embodiments according to the present disclosure. For example, any matter described in the documents cited in the present disclosure and any known technology related to the technical field related to the present disclosure can be combined with the embodiments according to the present disclosure. [[ID=�6]]

[0015] Furthermore, any matter disclosed by the applicant of the present application regarding the technology related to the present disclosure (for example, a matter freely selected from the disclosed matters in patent applications, papers, etc. made by the applicant of the present application) can be combined with the embodiments according to the present disclosure.

[0016] Any two or more of various exemplary aspects of the embodiments according to the present disclosure can be combined. Note that the combination of some technical matters may be the overall combination or partial combination of those technical matters.

[0017] One object of the embodiments according to the present disclosure is to improve the quality of medical data. The embodiments according to the present disclosure attempt to achieve this object by focusing on the consistency between medical data. More specifically, the embodiments according to the present disclosure improve the quality of the overall medical data (for example, the overall medical data regarding a certain patient) by evaluating the consistency between reference data (reference data, first medical data) and other data (target data, second medical data).

[0018] This "consistency" has various aspects. Examples of the consistency between data include the presence or absence of target data, the excess or deficiency of target data, the match / mismatch between reference data and target data, the necessity of target data based on reference data, the possibility of target data based on reference data, and the like.

[0019] The presence or absence of target data is, for example, the existence or non-existence of target data corresponding to the reference data. The excess or deficiency of target data is the excessive existence or deficiency of target data corresponding to the reference data. The match / mismatch between reference data and target data is the match or non-match between the reference data and the corresponding target data.

[0020] The necessity of target data based on reference data is similar to matters of "existence or absence" or "excess or deficiency," where the existence of reference data inevitably suggests the existence of specific target data. Examples of this necessity include: if a test order exists as reference data, then data corresponding to this test order (for example, data obtained from the test corresponding to that order (test data), test implementation information indicating that the test corresponding to that order was performed, etc.) must necessarily exist; conversely, if test data (test results) exists as reference data, then order information corresponding to the test performed to obtain this test data, and test implementation information indicating that the test performed to obtain this test data, must also exist.

[0021] The possibility of target data based on reference data is similar to matters of "presence or absence" or "excess or deficiency," where the existence of reference data indicating a certain possibility suggests the existence of specific target data. An example of this possibility is as follows: If a record (reference data) indicates that a candidate disease was identified from image or test data, then the disease name, corresponding test name, corresponding data, etc., may be recorded in other records (target data).

[0022] The types of data used to evaluate consistency include user-entered data (input data), data obtained by processing medical data (processed data), and data that indicates judgment criteria and action guidelines (guideline data).

[0023] Examples of input data include electronic medical record data created by physicians and image interpretation reports created by radiologists. Examples of processed data include data generated by processing (processing, analysis, etc.) examination data and image data, and data extracted from examination data and image data. Examples of guideline data include clinical practice guidelines and clinical pathways.

[0024] Thus, the embodiments relating to this disclosure improve the quality of medical data by evaluating the consistency between reference data and target data.

[0025] Reference data may include one or more of the following: first input data entered by healthcare professionals (e.g., electronic medical record data, surgical plans, surgical records, treatment plans, inpatient treatment plans, prescription data, test orders, medical accounting data, nursing records, types and model numbers of devices applied to patients (intraocular lenses, minimally invasive glaucoma surgery (MIGS) devices, artificial blood vessels, artificial organs, artificial retinas, etc.)), first test data obtained based on data obtained through medical tests (e.g., measurement data, imaging data, endoscopic data, biochemical test data, biopsy data, genetic test data), and pre-created standard data (e.g., clinical guidelines, clinical pathways, normative data).

[0026] Here, standard data includes information on standard medical practices (diagnosis, examination, treatment, surgery, drug prescription, hospital admission and discharge, rehabilitation, medical accounting, nursing, etc.). In other words, standard data contains standardized information on medical practices. The aforementioned guideline data, which standardizes judgment criteria and behavioral guidelines, can be treated as standard data.

[0027] On the other hand, the target data may include second input data entered by healthcare professionals (e.g., electronic medical record data, surgical plans, surgical records, treatment plans, inpatient treatment plans, prescription data, test orders, medical accounting data, nursing records, types and model numbers of devices applied to patients, etc.) and / or second test data obtained based on data obtained through medical tests (e.g., measurement data, imaging data, endoscopic data, biochemical test data, biopsy data, genetic test data, etc.).

[0028] Medical examinations performed to obtain the first and / or second examination data may include ophthalmic examinations. The examination data obtained by these ophthalmic examinations may include, for example, at least one of the following: data obtained by a tonometer (intraocular pressure data); data obtained by an optical coherence tomography (OCT) device (OCT data); data obtained by an ophthalmic imaging device (slit lamp microscope, fundus camera, scanning laser ophthalmoscope, ophthalmic surgical microscope, anterior segment observation device, posterior segment observation device, etc.) (image data); data obtained by a strabismus examination (strabismus data); data obtained by a heterophoria examination (heterophoria data); data obtained by a visual acuity examination Data (visual acuity data); data obtained from depth perception tests (depth perception data); data obtained from refraction measurements (refractive power data); data obtained from aberration measurements (ocular aberration data); data obtained from visual field tests (visual field data); data obtained from corneal morphology analysis (corneal morphology data); data obtained from fundus morphology analysis (fundus morphology data); data obtained from specular microscopes (corneal endothelial cell data); data obtained from axial length measurements (axial length data); data obtained from binocular vision function tests (binocular vision function data); data obtained from color vision tests (color vision data). Note that ophthalmic examination data are not limited to these types and may be any type. Furthermore, in fields other than ophthalmology, the first examination data and / or the second examination data may include any type of examination data that can be referenced in that field.

[0029] The embodiments relating to this disclosure can perform processing to resolve consistency issues when it is determined that there are problems with the consistency of medical data. For example, some embodiments can generate and add annotation information corresponding to consistency issues. Annotation information may include, for example, information to correct consistency issues, information to suggest corrections to consistency issues, additional information to resolve consistency issues, and information requesting the user to perform problem-solving work.

[0030] While the timing of performing consistency assessments between medical data can be arbitrary, several exemplary configurations allow for real-time consistency assessments.

[0031] For example, while a physician is entering information into an electronic medical record, consistency evaluation between medical data can be performed in real time. In other words, consistency evaluation between medical data can be performed in parallel with the electronic medical record creation process. Electronic medical records must contain a wide range of information sufficiently, completely, and accurately. However, regardless of the size of the medical institution, actual medical practice (especially outpatient settings) demands increased efficiency and throughput in medical care, and there is always a possibility of omissions or errors in electronic medical record entries. Utilizing the consistency evaluation between medical data described in this disclosure to automatically detect omissions and errors in electronic medical records is expected to contribute not only to increased efficiency in medical care but also to improved quality of medical care delivery. Furthermore, by performing consistency evaluation in parallel with the electronic medical record creation process (i.e., by performing consistency evaluation in real time), it is expected that efficiency in medical care and quality of medical care delivery can be further improved. Note that the real-time consistency evaluation is not limited to electronic medical record creation, but can apply to any task or process.

[0032] This disclosure describes embodiments of a medical data processing system, a medical data processing device, a medical data processing method, a program, and a recording medium, but the embodiments are not limited to these categories.

[0033] At least some of the functions of the elements described in the embodiments of this disclosure are implemented using a circuitry or processing circuitry. The circuitry or processing circuitry is configured and / or programmed to perform at least some of the disclosed functions and is a general-purpose processor, dedicated processor, integrated circuit, CPU (Central Processing Unit), GPU (Graphics Processing Unit), ASIC (Application Specific Integrated Circuit), programmable logic device (e.g., SPLD (Simple Programmable Logic Device), CPLD (Complex Programmable Logic Device), FPGA (Field Programmable Gate) This includes any array, conventional circuit configurations, and any combination thereof. A processor is considered a processing circuit configuration or circuit configuration, including transistors and / or other circuit configurations. In this disclosure, a circuit configuration, unit, means, or similar terms means hardware that performs at least a portion of the disclosed functions, or hardware programmed to perform at least a portion of the disclosed functions. The hardware may be the hardware described in the embodiments of this disclosure, or it may be known hardware programmed and / or configured to perform at least a portion of the described functions. If the hardware is a processor that can be considered a certain type of circuit configuration, then a circuit configuration, unit, means, or similar terms means a combination of hardware and software, the software being used to configure the hardware and / or processor. At least a portion of the functions of the elements described in the embodiments of this disclosure may be configured using artificial intelligence techniques such as machine learning.

[0034] <Medical Data Processing System> Several exemplary embodiments of the medical data processing system according to this embodiment will be described.

[0035] Figure 1 shows an example of an information system built within a medical institution. This information system is a network that includes multiple devices (computers, medical equipment, subsystems, etc.) connected via communication line 2000.

[0036] The information system in this example includes, as exemplary elements, at least one medical data processing system 1000, at least one hospital information system (HIS) 2100, at least one radiology information system (RIS) 2200, at least one picture archiving and communication system (PACS) 2300, at least one departmental system 2400, at least one examination device 2500, at least one medical professional terminal 2600, and so on.

[0037] The types of elements included in an information system for medical institutions are not limited to those shown in Figure 1; for example, a clinical laboratory information system (LIS) or a reporting system may also be included in the information system.

[0038] The medical data processing system 1000 is one exemplary embodiment of the medical data processing system according to the embodiment. The functions, configuration, operation, etc. of the medical data processing system 1000 will be described later.

[0039] The Hospital Information System 2100 is an information network that supports operations by connecting multiple departments within a medical institution. Examples of subsystems included in the Hospital Information System 2100 include, but are not limited to, automated reception systems, electronic medical record systems, order entry systems (ordering systems), medical accounting systems, admission and discharge management systems, pharmacy management systems, and appointment scheduling systems.

[0040] The Radiology Information System 2200 is a system for managing information related to examinations using various medical imaging devices, including radiation equipment. The Radiology Information System 2200 manages information by acquiring patient information, appointment information, and examination information from the Hospital Information System 2100. The Radiology Information System 2200 is built in accordance with standards such as Digital Imaging and Communications in Medicine (DICOM), which defines medical image formats and communication protocols between medical image handling devices.

[0041] Examples of medical imaging devices whose images are managed by the Radiology Information System 2200 include, but are not limited to, X-ray diagnostic equipment, X-ray computed tomography (X-ray) scanners, magnetic resonance imaging (MRI) scanners, positron emission tomography (POST) scanners, single-photon emission tomography (SMT) scanners, ultrasound diagnostic equipment, endoscopes, and ophthalmic imaging equipment.

[0042] The image storage and communication system 2300 has the function of receiving images from various medical imaging devices and saving them to a database, and the function of providing the saved images to terminals (such as medical professional terminals 2600).

[0043] Departmental System 2400 is an information system installed in each department of a medical institution. Examples of departments in a medical institution include, but are not limited to, clinical departments (various medical specialties), nursing departments, pharmacy departments, laboratory departments, radiology departments, nutrition departments, rehabilitation departments, administrative departments, information management departments, business planning departments, and medical consultation departments.

[0044] The 2500 testing device is a device for performing medical tests. Examples of the 2500 testing device include, but are not limited to, physical examination devices, specimen testing devices, culture testing devices, functional testing devices, medical imaging devices, electrophysiology testing devices, respiratory physiology testing devices, endoscope devices, sensory organ testing devices, motor function testing devices, pathology diagnostic devices, psychological testing devices, and genetic testing devices.

[0045] The Medical Professional Terminal 2600 is a computer terminal used by medical professionals. Examples of medical professionals who use the Medical Professional Terminal 2600 include, but are not limited to, physicians, dentists, pharmacists, nurses, midwives, public health nurses, certified psychologists, registered dietitians, clinical laboratory technologists, radiological technologists, clinical engineers, dental hygienists, physical therapists, occupational therapists, prosthetists and orthotists, dental technicians, paramedics, speech-language pathologists, and orthoptists.

[0046] Next, the medical data processing system 1000 will be described in detail. The medical data processing system 1000 is an information processing system configured to perform consistency evaluations between medical data. An example of the configuration of the medical data processing system 1000 is shown in Figure 2.

[0047] The medical data processing system 1000 in this example includes a medical data acquisition unit 1100, an evaluation unit 1200, an annotation unit 1300, an output unit 1400, and a user interface (UI) 1500.

[0048] The elements included in the medical data processing system according to the embodiment are not limited to those shown in Figure 2. For example, some exemplary embodiments of the medical data processing system may include only at least a portion of the evaluation unit 1200 or elements having equivalent functionality from among the elements shown in Figure 2. Also, some exemplary embodiments of the medical data processing system may include at least a portion of the evaluation unit 1200 or equivalent functionality and one or more of the elements shown in Figure 2 other than the evaluation unit 1200. Furthermore, some exemplary embodiments of the medical data processing system may include at least a portion of the evaluation unit 1200 or equivalent functionality and any elements not shown in Figure 2.

[0049] Now, let's explain each element of the medical data processing system 1000 in this example. First, the medical data acquisition unit 1100 acquires medical data to which consistency evaluation is applied. The medical data acquisition unit 1100 acquires medical data from one or more of the following, for example, the hospital information system 2100, the radiology department information system 2200, the image storage and communication system 2300, the department system 2400, the examination device 2500, and the medical professional terminal 2600.

[0050] In some exemplary embodiments, the medical data acquisition unit 1100 can acquire predetermined medical data from these data sources by requesting the provision of predetermined medical data from those data sources via a communication line 2000. The medical data is specified, for example, by a patient identifier (such as a patient ID), a date, a medical procedure identifier (such as an examination ID, image ID, surgery ID, prescription ID, or physician accounting ID), or a combination thereof. The data source that receives the request sends the medical data corresponding to this request to the medical data processing system 1000.

[0051] In some exemplary embodiments, a user (a healthcare professional such as a physician) can use a healthcare professional terminal 2600 to specify the medical data to which they wish to apply consistency evaluation, and send the specified medical information to a medical data processing system 1000.

[0052] In some exemplary embodiments, the healthcare professional terminal 2600 may be configured to automatically perform the processes of extracting predetermined types of medical data (for example, medical data entered, processed, edited, or deleted by the user) from the medical data used by the user, and transmitting the extracted medical data to the medical data processing system 1000.

[0053] In some exemplary embodiments, the medical professional terminal 2600 may be configured to automatically perform the following processes: extracting predetermined types of medical data (for example, medical data entered, processed, edited, or deleted by the user) from the medical data used by the user; presenting the extracted medical data to the user; receiving the user's actions on the presented medical data; and transmitting the medical data specified by the user from the presented medical data to the medical data processing system 1000.

[0054] The functions of the medical data acquisition unit 1100 are realized through the cooperation of software, such as a medical data acquisition program, and hardware, such as a processor. The medical data acquisition unit 1100 operates under the control of, for example, a control processor (control unit) not shown. This control processor is configured to execute control of the medical data acquisition unit 1100 based on, for example, a control program not shown.

[0055] Furthermore, the medical data acquisition unit 1100 includes a communication interface for data communication with other devices via the communication line 2000. In some exemplary embodiments, the medical data acquisition unit 1100 may include a data reader for reading data recorded on a recording medium, and / or a scanner or optical character recognition (OCR) software for reading information recorded on paper sheets. More generally, the medical data acquisition unit 1100 may have any functions, hardware, software, etc., that can be used to acquire medical data from an external source.

[0056] The medical data acquired by the medical data acquisition unit 1100 will now be described. The medical data acquisition unit 1100 acquires medical data (reference data) to be used as a standard in the consistency evaluation performed by the evaluation unit 1200, and medical data (target data) to be used as a comparison target for this reference data. In some exemplary embodiments, the medical data processing system 1000 (for example, the evaluation unit 1200) may select the reference data and target data from among the various data acquired by the medical data acquisition unit 1100.

[0057] As mentioned above, the types of medical data used for consistency evaluation include data entered by users (input data), data obtained by processing medical data (processed data), and data that shows judgment criteria and action guidelines (guideline data).

[0058] As also mentioned above, the reference data in this embodiment includes one or more of the following: input data entered by healthcare professionals (reference input data, first input data), test data obtained based on data obtained by medical tests (reference test data, first test data), and pre-created standard data. Furthermore, the target data in this embodiment includes input data entered by healthcare professionals (target input data, second input data), and / or test data obtained based on data obtained by medical tests (target test data, second test data). Figure 3 shows the combinations of reference data and target data that are compared in the consistency evaluation of this embodiment.

[0059] In some exemplary embodiments, the medical data acquisition unit 1100 may include a test data acquisition unit 1110. If the medical data acquisition unit 1100 includes a test data acquisition unit 1110, the medical data acquisition unit 1100 can acquire data obtained by medical tests from an external source and generate test data (reference test data or target test data) by processing this data using the test data acquisition unit 1110.

[0060] In other words, while a medical data acquisition unit 1100 without an inspection data acquisition unit 1110 functions to acquire inspection data (reference data, target data) provided to the evaluation unit 1200 from an external source, a medical data acquisition unit 1100 with an inspection data acquisition unit 1110 can acquire inspection data (reference data, target data) provided to the evaluation unit 1200 from an external source, and can also generate inspection data (reference data, target data) provided to the evaluation unit 1200 based on data acquired from an external source.

[0061] If the medical data acquisition unit 1100 includes a test data acquisition unit 1110, and the reference data includes reference test data, then this reference test data may be acquired (generated) by the test data acquisition unit 1110, and / or, if the target data includes target test data, then this target test data may be acquired (generated) by the test data acquisition unit 1110.

[0062] The test data acquisition unit 1110 may be configured to perform processes for generating test data from data generated by medical tests and acquired by the medical data acquisition unit 1100, such as an extraction process to extract test data from the data, a processing process to process the data to generate test data, a calculation process to calculate test data by applying calculations to the data, and an inference process to obtain test data by performing inferences based on the data. These processes are performed based on a predetermined algorithm. The inference process may be performed using artificial intelligence technology. For example, the inference process can be performed using a machine learning model trained to receive a specific type of data as input and output test data. This machine learning is performed, for example, using training data that includes a set of pairs of a specific type of data and test data.

[0063] When the medical data acquisition unit 1100 includes an examination data acquisition unit 1110, and the data generated by a medical examination and acquired by the medical data acquisition unit 1100 includes medical image data, the examination data acquisition unit 1110 can generate image information and / or numerical information based on this medical image data. As a process for generating image information and / or numerical information from medical image data, the examination data acquisition unit 1110 may be configured to perform, for example, a process of analyzing medical image data using a predetermined analysis algorithm, a process of comparing medical image data with predetermined medical image data using a predetermined comparison algorithm, and an inference process using a machine learning model trained to receive medical image data as input and output image information and / or numerical information. This machine learning is performed, for example, using training data that includes a set of pairs of medical image data and examination data (image information and / or numerical information).

[0064] For example, the examination data acquisition unit 1110 can generate layer data (e.g., retinal thickness data, choroidal thickness data, comparison results with standard databases (normal eye database, diseased eye database, etc.)) from fundus OCT images acquired by applying an OCT scan to the patient's eye.

[0065] As another example, the test data acquisition unit 1110 can acquire candidate disease names (diagnoses) from electronic medical records and image interpretation reports, and based on the extracted candidate disease names, it can extract and / or generate test data (reference test data and / or target test data) from various medical data of the patient. The extraction and / or generation of test data based on candidate disease names may be performed with and / or without machine learning techniques, for example. If machine learning techniques are used, the test data acquisition unit 1110 may be configured to perform inference processing using, for example, a machine learning model trained using a large number of documents (medical papers, etc.) and trained to receive a disease name as input and output the type of test (type of test data). As a process that does not use machine learning techniques, the test data acquisition unit 1110 may be configured to identify the type of test corresponding to a candidate disease name acquired from electronic medical records and image interpretation reports, for example, by using a table that shows the correspondence between a large number of disease names and a large number of test types.

[0066] In some exemplary embodiments, a computer (inspection data generation device) having a configuration similar to that of the inspection data acquisition unit 1110 may be provided separately from the medical data processing system. In such embodiments, the medical data processing system can acquire the inspection data generated by the inspection data generation device and provide it for consistency evaluation.

[0067] The evaluation unit 1200 is configured to perform an evaluation of the consistency between the reference data and the target data.

[0068] The functions of the evaluation unit 1200 are realized through the cooperation of software, such as a consistency evaluation program, and hardware, such as a processor. The evaluation unit 1200 operates under the control of a processor (control unit), for example, not shown. This processor is configured to execute control of the evaluation unit 1200 by a control program, for example, not shown.

[0069] As mentioned above, consistency between reference data and target data can manifest in various ways, including the presence or absence of target data, surplus or deficiency of target data, agreement or disagreement between reference data and target data, the necessity of target data based on reference data, and the possibility of target data based on reference data. Several examples of consistency and several examples of consistency evaluation are described below.

[0070] In some exemplary embodiments, the evaluation unit 1200 may include a determination unit 1210. If the evaluation unit 1200 includes a determination unit 1210, the evaluation unit 1200 can obtain information (determination result) obtained by the determination unit 1210 as the result of consistency evaluation, and / or process the information obtained by the determination unit 1210 to obtain the result of consistency evaluation.

[0071] The functions of the determination unit 1210 are realized through the cooperation of software, such as a determination program, and hardware, such as a processor. The determination unit 1210 operates under the control of a processor (control unit), for example, not shown. This processor is configured to execute control of the determination unit 1210 by a control program, for example, not shown.

[0072] Several examples of the functions and configuration of the determination unit 1210 are described below.

[0073] The determination unit 1210 may be configured to determine whether the information contained in the reference data is included in the target data (first determination unit). This makes it possible to detect the omission of information that should be included in the target data.

[0074] For example, if a radiographic interpretation report created based on a patient's medical images is adopted as reference data, and the patient's electronic medical record created by referring to this radiographic interpretation report is adopted as target data, the determination unit 1210 can perform the process of extracting candidate disease names from the radiographic interpretation report and the process of searching for the extracted candidate disease names in the electronic medical record. If a candidate disease name is found in the electronic medical record, the determination unit 1210 determines that there is no omission in entering the candidate disease name based on the radiographic interpretation report into the electronic medical record. Conversely, if a candidate disease name is not found in the electronic medical record, the determination unit 1210 determines that there is an omission in entering the candidate disease name based on the radiographic interpretation report into the electronic medical record.

[0075] As another example, if the clinical practice guidelines for a disease state that a second test should be performed based on the results obtained from a first test performed on a patient suspected of having a certain disease, the determination unit 1210 can determine whether or not information regarding the second test (for example, an order for the second test) is recorded in the electronic medical record of the patient with the disease by comparing the clinical practice guidelines (reference data) with the electronic medical record of the patient with the disease (target data). This makes it possible to detect omissions in the electronic medical record and prevent the second test from being missed.

[0076] The determination unit 1210 may be configured to determine whether the information contained in the reference data matches the information contained in the target data (second determination unit). This makes it possible to detect discrepancies in content between the information contained in the reference data and the information contained in the target data.

[0077] For example, when a certain test is performed on a patient, the determination unit 1210 can determine whether the type of test indicated in the test order matches the type of test recorded in the electronic medical record by comparing the test order (reference data) for that test with the patient's electronic medical record (target data) created after the test. This makes it possible to detect omissions or errors in the electronic medical record.

[0078] The determination unit 1210 may be configured to determine whether the target data contains information corresponding to the information contained in the reference data (third determination unit). This makes it possible to determine whether the target data contains content suggested by the reference data.

[0079] For example, if a certain test is performed on a patient, a test order for that test exists, and the existence of this test order suggests that test data obtained from that test exists, as well as test implementation information indicating that the test was performed. In other words, if a test order exists, then at any date and time after the test date and time indicated by this test order, the patient's medical data (e.g., electronic medical record) will inevitably contain test data and test implementation information. The determination unit 1210 can determine whether or not the test data and / or test implementation information corresponding to the test indicated by this test order is included by comparing this test order (reference data) with the patient's electronic medical record (target data). This makes it possible to detect omissions in the electronic medical record.

[0080] Conversely, if a patient's medical data includes test data (or test implementation information) for a particular test, then a test order for that test should exist. The determination unit 1210 can determine whether or not a corresponding test order exists by comparing this test data (reference data) with the test order associated with the patient. This makes it possible to detect missing test orders or errors in the test flow.

[0081] As another example, if a candidate disease is identified from a patient's medical image, it is assumed that the name of the identified candidate disease, the name of the test corresponding to this candidate disease, and the test data obtained from that test are included in the patient's medical data (such as an electronic medical record). In other words, if a candidate disease is identified, there is a possibility that data corresponding to this candidate disease is included in the patient's medical data. The determination unit 1210 can determine whether or not information corresponding to this candidate disease is included in the patient's medical data by comparing medical data containing the name of the candidate disease identified from the medical image (reference data: for example, the patient's image interpretation report) with various medical data of the patient (target data: for example, the patient's electronic medical record). This makes it possible to suggest the possibility of omissions or errors in the medical data.

[0082] Medical data comes in various forms, and relationships (causal relationships, correlations, etc.) may exist between different types of medical data. In such cases, it is desirable to consider the combination of these medical data when performing consistency assessments.

[0083] For example, certain medical procedures can affect the results of certain tests. Specifically, surgery can cause elevated blood pressure, or prolonged bed rest after surgery can lead to a hypercoagulable state (tendency to form blood clots). Furthermore, side effects of administered medications can impair kidney or liver function. Therefore, it is important to identify whether the test data obtained for a particular patient has been influenced by a medical procedure. However, given the extremely large number of possible combinations of medical procedures and test results, and the wide-ranging impact of medical procedures on test results, it cannot be said that there is no possibility of overlooking the relationship between medical procedures and test results.

[0084] Therefore, several exemplary embodiments are configured to detect omissions and errors in electronic medical records by considering the combination of medical procedures performed on a patient and the patient's test data. Furthermore, by presenting users with information about the combination of medical procedures performed on a patient and the patient's test data, and by presenting users with information about the impact of medical procedures on test data, it becomes possible to prevent omissions and errors in electronic medical records. This improves the quality of consistency evaluation between medical data. In addition, by providing this information to patients, it is thought that explanations can be appropriately provided to patients who have questions or concerns about their test data.

[0085] To achieve such effects, in some exemplary embodiments, the reference data includes a group of reference data (first data group) consisting of two or more data points, and / or the target data includes a group of target data (second data group) consisting of two or more data points. Furthermore, the evaluation unit is configured to perform a consistency evaluation based on the reference data group and / or the target data group.

[0086] Let's explain one specific example. In some exemplary embodiments, the reference data set includes standard range information for a predetermined test item and relational information indicating the relationship between this predetermined test item and a predetermined medical procedure. In this example, the target data set includes test data obtained by applying the predetermined test item indicated in the standard range information to a patient, and medical procedure information indicating the medical procedure performed on this patient. The standard range information and relational information of the reference data set are, for example, pre-created as standard data. The test data and medical procedure information of the target data set are, for example, obtained from the patient's electronic medical record data.

[0087] In this example, the evaluation unit 1200 first refers to the relationship information of the reference data group to determine whether there is a relationship between the inspection of predetermined items shown in the standard range information of the reference data group and the medical procedures shown in the medical procedure information of the target data group. This determination process includes, for example, a process of comparing the medical procedures shown in the medical procedure information of the target data group with the medical procedures shown in the relationship information of the reference data group. If any medical procedure included in the medical procedure information matches any medical procedure shown in the relationship information, it is determined that there is a relationship between the inspection of predetermined items shown in the standard range information of the reference data group and the medical procedures shown in the medical procedure information of the target data group. Otherwise, it is determined that there is no relationship between the inspection of predetermined items shown in the standard range information of the reference data group and the medical procedures shown in the medical procedure information of the target data group.

[0088] If the above determination process determines that there is "no relationship" between the examination of a predetermined item indicated in the standard range information of the reference data group and the medical procedure indicated in the medical procedure information of the target data group, the evaluation unit 1200 in this example compares the standard range information of the reference data group with the examination data of the target data group. This comparison process determines, for example, whether the value indicated by the examination data of the target data group falls within the range indicated by the standard range information of the reference data group (e.g., normal range, abnormal range). Based on the results obtained from this comparison process, the evaluation unit 1200 in this example evaluates the consistency between the reference data and the target data. This consistency evaluation may, for example, evaluate the consistency (omissions, incorrect entries, etc.) between the contents of the electronic medical record of the patient included in the target data and the comparison result. As a specific example, this consistency evaluation may determine whether the comparison result is reflected in the contents of the electronic medical record of the patient.

[0089] On the other hand, if the above determination process determines that there is a "relationship" between the examination of a predetermined item indicated in the standard range information of the reference data group and the medical procedure indicated in the medical procedure information of the target data group, the evaluation unit 1200 in this example first compares the standard range information of the reference data group with the examination data of the target data group, similar to the case where there is "no relationship" described above. Next, the evaluation unit 1200 in this example can process the results obtained from this comparison process based on the relationship information of the reference data group. This processing includes, for example, a process to change the comparison result based on the relationship information, a process to attach the information indicated in the relationship information (information showing the relationship between the examination and the medical procedure) to the comparison result, a process to flag the comparison result, and a process to change the display manner of the comparison result (for example, highlighting). Based on the results obtained from this processing, the evaluation unit 1200 in this example evaluates the consistency between the reference data and the target data. This consistency evaluation may, for example, assess the consistency (omissions, incorrect entries, etc.) between the contents of the patient's electronic medical record included in the target data and the results obtained from the processing. Specifically, this consistency evaluation may determine whether or not the results obtained from the processing are reflected in the contents of the patient's electronic medical record.

[0090] With this configuration, it is possible to evaluate the consistency between medical data, taking into account the impact of medical procedures performed on the patient on the test results. Furthermore, if the medical procedures performed on the patient do not affect the test results, the same consistency evaluation can be performed as in other configurations.

[0091] In some exemplary embodiments, a predetermined medical procedure indicated in the relational information of the reference data group may include a predetermined surgery (surgical name, surgical procedure, type of surgery, surgical identifier, etc.), and the medical procedure information of the target data group may include surgical information (surgical name, surgical procedure, type of surgery, surgical identifier, etc.) indicating the surgery performed on the patient. The evaluation unit 1200 in this embodiment can perform a determination process based on such relational information to determine whether or not there is a relationship between the examination of a predetermined item indicated in the standard range information of the reference data group and the surgery indicated in the surgical information of the medical procedure information of the target data group. Furthermore, based on the result of this determination process, the evaluation unit 1200 in this embodiment can perform the same processing as described above for the case where there is a relationship, or the same processing as described above for the case where there is no relationship. According to this embodiment, it is possible to evaluate the consistency between medical data while considering the impact that the surgery performed on the patient has on the examination results. If the surgery performed on the patient does not have an impact on the examination results, the same consistency evaluation as in other embodiments can be performed.

[0092] In some exemplary embodiments, a predetermined medical procedure indicated in the relational information of the reference data group may include the administration of a predetermined drug (drug name, drug identifier, etc.), and the medical procedure information of the target data group may include drug information (drug name, drug identifier, drug history, etc.) indicating the drug prescribed to the patient. The evaluation unit 1200 in this embodiment can perform a determination process based on such relational information to determine whether or not there is a relationship between the examination of a predetermined item indicated in the standard range information of the reference data group and the drug indicated in the drug information of the medical procedure information of the target data group. Furthermore, based on the result of this determination process, the evaluation unit 1200 in this embodiment can perform the same processing as described above for the case where there is a relationship, or the same processing as described above for the case where there is no relationship. According to this embodiment, it is possible to evaluate the consistency between medical data while considering the effect that the drug administered to the patient has on the test results. If the drug administered to the patient does not affect the test results, the same consistency evaluation as in other embodiments can be performed.

[0093] Let's explain an example in ophthalmology. In this example, the standard range information of the reference data group includes information indicating the normal and / or abnormal range of intraocular pressure (IOP) values, and the relationship information of the reference data group includes relationship information indicating the relationship between IOP tests and drugs that may affect IOP values ​​(e.g., corticosteroids). In addition, the target data group in this example includes the patient's IOP data and drug information indicating the drugs administered to this patient. The evaluation unit 1200 determines whether there is a relationship between the IOP test and the drugs indicated in this drug information. This allows us to determine whether the drugs administered to the patient affect IOP values. If the drugs administered to the patient do not affect IOP values, the evaluation unit 1200 can evaluate whether the results of the IOP test are correctly reflected in the electronic medical record. On the other hand, if the drugs administered to the patient do affect IOP values, the evaluation unit 1200 can, for example, apply the aforementioned processing to the IOP data in the target data group and evaluate the electronic medical record based on the information obtained therefrom. In this example, if none of the medications administered to the patient affect the intraocular pressure (IOP) value, the IOP value obtained from the IOP test can be compared with standard range information, allowing for a normal IOP determination. However, if any of the medications administered to the patient does affect the IOP value, this fact should be communicated to healthcare professionals and the patient, and the IOP value should be determined taking this effect into consideration.

[0094] In some exemplary embodiments, the evaluation unit 1200 may be configured to determine whether the site to be treated (such as surgery) coincides with the planned site.

[0095] In this example, the reference data includes patient treatment site information as input data entered by healthcare professionals (reference input data, first input data). Treatment site information is information indicating the area to be treated and is included, for example, in medical data managed by the hospital information system 2100 (e.g., surgical plan, treatment plan, medical treatment plan, electronic medical record, etc.).

[0096] Furthermore, in this example, the target data includes examination data (target examination data, second examination data) obtained based on data obtained through medical examinations, as well as image data obtained during patient treatment (treatment-time image data). Treatment-time image data is managed, for example, by the image storage and communication system 2300. The hospital information system 2100, departmental system 2400, examination device 2500, medical professional terminal 2600, or other devices may also temporarily hold treatment-time image data.

[0097] In this example, the evaluation unit 1200 first identifies the treatment target area of ​​the patient based on the treatment image data. This identification process may include, for example, image analysis of the treatment image data and / or inference using a machine learning model. This image analysis includes, for example, image segmentation. Furthermore, this machine learning model is constructed, for example, by machine learning using a set of medical images with annotation information indicating the treatment target area, and functions to receive medical images (treatment image data) as input and output the treatment target area.

[0098] Furthermore, the evaluation unit 1200 in this example performs a consistency evaluation between the reference data and the target data by comparing the treatment target site identified by the above identification process with the treatment target site information included in the reference data. This comparison process includes, for example, determining whether the treatment target site identified from the treatment image data matches the site indicated in the treatment target site information.

[0099] In this example, the evaluation unit 1200 can perform such treatment target site matching while the treatment is being performed. In other words, the treatment target site matching process in this example may be a real-time process. For example, while performing treatment on a patient, the patient can be photographed using an imaging device (not shown) to acquire treatment image data, and the acquired treatment image data can be input to the evaluation unit 1200, allowing the evaluation unit 1200 to perform treatment target site matching in real time. Such real-time processing allows the planned treatment target site to be presented to the medical professional during treatment, and a warning can be issued if treatment is attempted on a site different from the planned treatment target site.

[0100] Furthermore, the evaluation unit 1200 in this example can perform a verification of the treatment site after the patient's treatment is completed. In other words, the verification process of the treatment site in this example may be a post-treatment process. For example, the evaluation unit 1200 can perform a verification of the treatment site by receiving treatment image data taken during treatment using an imaging device (not shown). Such post-treatment processing makes it possible to determine, for example, whether the planned treatment site was treated correctly, and to evaluate the contents of the patient's electronic medical record and surgical record.

[0101] In some exemplary embodiments, the evaluation unit 1200 may be configured to determine whether the instruments used in the treatment (such as surgery) are consistent with those planned. Examples of treatment instruments include intraocular lenses, minimally invasive glaucoma surgery (MIGS) devices, artificial blood vessels, artificial organs, and artificial retinas. Each treatment instrument is assigned a unique identifier (such as a model number).

[0102] In this example, the reference data includes patient treatment device information as input data entered by healthcare professionals (reference input data, first input data). The treatment device information is information about the treatment device and includes, for example, a treatment device identifier. The treatment device information is included in medical data managed by the hospital information system 2100 (for example, surgical plans, treatment plans, medical treatment plans, electronic medical records, etc.).

[0103] Furthermore, in this example, the target data includes data obtained during patient treatment (treatment data) as examination data (target examination data, second examination data) obtained based on data obtained through medical examinations. Treatment data includes, for example, image data obtained by photographing treatment instruments during treatment (treatment instrument image data) and data obtained by reading identifiers (such as barcodes) attached to treatment instruments during treatment (treatment instrument identifier data). Treatment data is managed by, for example, the hospital information system 2100, the image storage and communication system 2300, etc. It should also be noted that the department system 2400, the examination device 2500, the medical professional terminal 2600, or other devices may at least temporarily hold the treatment data.

[0104] In this example, the evaluation unit 1200 first identifies the instruments used in the treatment of the patient (treatment instruments used or employed in the treatment) based on the treatment data. If the treatment data includes treatment instrument identifier data, this identification process may include identifying the instruments used from this treatment instrument identifier data. If the treatment data includes treatment image data, this identification process may include image analysis of the treatment image data and / or inference using a machine learning model. This image analysis includes, for example, the process of applying image segmentation to the treatment image data to detect images of treatment instruments, and the process of analyzing the detected images of treatment instruments to identify the treatment instruments (instruments used). Furthermore, this machine learning model is constructed, for example, by machine learning using a set of images including images of treatment instruments with annotation information indicating the identifier of the treatment instrument (type, model number, etc.), and functions to receive treatment image data as input and output the identifier of the treatment instrument.

[0105] Furthermore, the evaluation unit 1200 in this example performs a consistency evaluation between the reference data and the target data by comparing the instruments used, which have been identified through the above identification process, with the treatment instrument information included in the reference data. This comparison process includes, for example, determining whether the instruments used, which have been identified from the treatment data, match the treatment instruments indicated in the treatment instrument information.

[0106] The evaluation unit 1200 in this example can perform such verification of treatment instruments while the treatment is being performed. In other words, the verification process of treatment instruments in this example may be a real-time process. For example, while treating a patient, the system may read the identifier attached to the treatment instrument using an identifier reader (not shown) to obtain treatment instrument identifier data, or it may take a picture of the patient using an imaging device (not shown) to obtain treatment image data. Furthermore, by inputting the acquired treatment data (treatment instrument identifier data, treatment image data, etc.) into the evaluation unit 1200, the evaluation unit 1200 can perform verification of treatment instruments in real time. Such real-time processing allows the type of treatment instrument planned to be presented to the medical professional during treatment, and also allows a warning to be issued if a treatment instrument different from the planned one is about to be used.

[0107] Furthermore, the evaluation unit 1200 in this example can perform a verification of treatment instruments after the patient's treatment is completed. In other words, the verification process of treatment instruments in this example may be a post-treatment process. For example, the evaluation unit 1200 can perform a verification of treatment instruments based on treatment data acquired during treatment. Such post-treatment processing can, for example, determine whether the treatment instruments were used as planned, and can also evaluate the contents of the patient's electronic medical record and surgical records.

[0108] In some exemplary embodiments, the evaluation unit 1200 may be configured to perform consistency assessments between medical data using a machine learning model. Several examples of such evaluation units 1200 are described below.

[0109] In some exemplary embodiments, the evaluation unit 1200 includes a machine learning model (first machine learning model) constructed by machine learning using training data that includes data of the same type as the reference data (first type of data) and data of the same type as the target data (second type of data). This machine learning model (inference model) receives input of at least pairs of reference data and target data and functions to output information (consistency evaluation information) indicating the result of an evaluation of the consistency between the reference data and the target data.

[0110] The evaluation unit 1200A shown in Figure 4 is one example configuration of the evaluation unit 1200 in Figure 2. In this example, the evaluation unit 1200A performs consistency evaluation using an inference model (machine learning model, mathematical model, trained model) 1220. The inference model 1220 includes a neural network 1221 that is trained by machine learning to receive at least pairs of reference data and target data as input and output consistency evaluation information about these data pairs.

[0111] The inference model 1220 may be located inside the evaluation unit 1200 as shown in Figure 4, or it may be located in another part of the medical data processing system 1000 (for example, in a storage device not shown), or it may be located outside the medical data processing system 1000. As an example of the latter, the inference model 1220 may be located inside a computer or storage device accessible by the medical data processing system 1000.

[0112] The device for constructing the inference model 1220 (inference model construction device) may be provided in the medical data processing system 1000, in peripheral equipment (such as a computer) of the medical data processing system 1000, or in another computer.

[0113] The inference model construction device 1600 shown in Figure 5 is an example of a device for constructing an inference model 1220, and includes a learning processing unit 1610 and a neural network 1620.

[0114] The structure and configuration of neural network 1620 are determined based on the types of input data (first type data, second type data), the types of output information (consistency evaluation information), and so on. For example, if the input data includes image data, neural network 1620 typically includes a convolutional neural network (CNN). Figure 5, code 1630, shows an example of the structure of this convolutional neural network.

[0115] The input layer of the convolutional neural network 1630 receives an image as input. Behind the input layer, multiple pairs of convolutional layers and pooling layers are arranged. In the example shown in Figure 5, there are three pairs of convolutional layers and pooling layers, but the number of pairs can be arbitrary.

[0116] In a convolutional layer, a convolution operation is performed to extract features (such as contours) from an image. A convolution operation is a sum-of-products operation of a filter function (weight coefficients, filter kernel) of the same dimension as the input image. In a convolutional layer, the convolution operation is applied to multiple parts of the input image. More specifically, in a convolutional layer, the value of each pixel in the part of the image to which the filter function is applied is multiplied by the value of the filter function corresponding to that pixel (weight), and the product is calculated. The sum of these products is then calculated across multiple pixels in this part of the image. The resulting sum-of-products is then assigned to the corresponding pixels in the output image. By performing the sum-of-products operation while moving the part of the image to which the filter function is applied, the result of the convolution operation on the entire input image is obtained. This type of convolution operation yields many images in which various features have been extracted using a large number of weight coefficients. In other words, many filtered images such as smoothed images and edge images are obtained. The many images generated by the convolutional layer are called feature maps.

[0117] In the pooling layer, the feature map generated by the preceding convolutional layer is compressed (e.g., data decimation). More specifically, the pooling layer calculates statistical values ​​for predetermined neighboring pixels of a pixel of interest within the feature map at predetermined pixel intervals, and outputs an image with dimensions smaller than the input feature map. The statistical values ​​applied to the pooling operation are, for example, the maximum value (max pooling) or the average value (average pooling). The pixel interval applied to the pooling operation is called the stride.

[0118] Convolutional neural networks can extract many features from an input image by processing it using multiple pairs of convolutional and pooling layers.

[0119] A fully connected layer is provided behind the last pair of convolutional and pooling layers. In the example shown in Figure 5, there are two fully connected layers, but the number of fully connected layers can be arbitrary. The fully connected layer uses the features compressed by the combination of convolution and pooling to perform processes such as image classification, image segmentation, and regression. An output layer is provided behind the last fully connected layer to provide the output result.

[0120] In some exemplary embodiments, the convolutional neural network may not include fully connected layers (e.g., a fully convolutional network (FCN)) and / or may include support vector machines or recurrent neural networks (RNNs). Furthermore, the machine learning applied to the neural network 1620 may include transfer learning. That is, the neural network 1620 may include a neural network that has already been trained using other training data and whose parameters have been tuned. The inference model building device 1600 (learning processing unit 1610) may also be configured to allow fine-tuning of the trained neural network (neural network 1620). The neural network 1620 may be constructed using a known open-source neural network architecture.

[0121] The learning processing unit 1610 applies machine learning using the training data 1670 created in the manner described later to the neural network 1620. If the neural network 1620 includes a convolutional neural network, the parameters adjusted by the learning processing unit 1610 include, for example, the filter coefficients of the convolutional layer and the connection weights and offsets of the fully connected layer.

[0122] The types of data included in the training data 1670 are determined based on the types of input data (Type 1 data, Type 2 data), the types of output information (consistency evaluation information), etc. The input data may be pairs of reference data of any of the aforementioned types and target data of any of the aforementioned types. The types of data included in the training data 1670 may be arbitrary and may include, for example, data created by healthcare professionals, data entered by healthcare professionals, data acquired by medical devices (such as the testing device 2500), data generated by medical devices, data created by processing any of these types of data, data generated by computers, pseudo-data, etc. Furthermore, the number of data items included in the training data 1670 may be increased by using techniques such as data augmentation.

[0123] The training method (machine learning method) for constructing the neural network 1221 may be, for example, supervised learning, but is not limited to this. In some exemplary embodiments, in addition to supervised learning, or in place of supervised learning, any known method such as unsupervised learning, reinforcement learning, semi-supervised learning, transduction, or multitask learning may be used in machine learning to construct the neural network 1221.

[0124] To avoid processing being concentrated on specific units of the neural network 1221, the learning processing unit 1610 may randomly select and disable some units of the neural network 1620 and perform machine learning using the remaining units (dropout).

[0125] The methods used to construct inference models are not limited to those shown here. In some exemplary embodiments, any known method, such as support vector machines, Bayesian classifiers, boosting, k-means algorithms, kernel density estimation, principal component analysis, independent component analysis, self-organizing maps, random forests, and generative adversarial networks (GANs), can be used to construct inference models.

[0126] Several examples of how to create training data 1670 are described below. In the example shown in Figure 5, a data source 1640, a data acquisition device 1650, and an annotation device 1660 are used to create the training data 1670. The number of data sources 1640 used in this embodiment may be one or any number. The same applies to the data acquisition device 1650 and the annotation device 1660.

[0127] Data source 1640 is the source of data for creating training data 1670. The type and number of data sources 1640 are arbitrary. Data sources 1640 may include, for example, a hospital information system 2100, a radiology information system 2200, an image storage and communication system 2300, a department system 2400, an examination device 2500, a medical professional terminal 2600, other devices and / or systems installed within the medical institution, devices and / or systems installed outside the medical institution, a recording medium, etc.

[0128] The data acquisition device 1650 includes a computer that has the function of collecting data from the data source 1640 to create training data 1670. In some exemplary embodiments, the data acquisition device 1650 may be configured to collect data from the data source 1640 of the same type as the data processed by the evaluation unit 1200A. That is, the data acquisition device 1650 may be configured to collect data of the same type as the reference data (first type of data) and data of the same type as the target data (second type of data) from the data source 1640. The set of data collected by the data acquisition device 1650 (dataset) is sent to the annotation device 1660.

[0129] The annotation device 1660 is used for annotation. This annotation is, for example, the process of attaching metadata (labels, tags) to the data collected by the data acquisition device 1650. These labels are information equivalent to the integrity evaluation information described above, that is, information of the same type as the integrity evaluation information. Annotation may be performed on each data item in the dataset collected by the data acquisition device 1650, or it may be performed only on some of the data items in this dataset.

[0130] Annotation work (manual annotation) is generally performed by physicians. The annotation device 1660 used for annotation work is equipped with hardware and software for performing annotation work. This hardware includes, for example, a computer, as well as user interface hardware (e.g., display devices, operating devices, input devices, etc.). The software includes, for example, a graphical user interface (GUI) application for performing processing for annotation work (e.g., display, acceptance of operations, acceptance of input, generation of annotation information, association of annotation information with data, etc.).

[0131] For example, a physician uses the annotation device 1660 used for annotation work to perform operations to display data, to determine the content of annotation information by referring to the displayed data, and to input the determined content of annotation information. Furthermore, the annotation device 1660 in this example performs a process to associate the annotation information input by the physician with the corresponding data.

[0132] Annotation processing (automatic annotation) is performed by a computer. The annotation device 1660 used for annotation processing comprises hardware and software for performing annotation processing. This hardware includes at least a computer. The software includes, for example, applications for performing annotation processing (e.g., data analysis, generation of annotation information, association of annotation information with data, etc.).

[0133] The annotation device 1660 may have a function for verifying the annotation results (e.g., annotation information, pairs of annotation information and data, etc.) obtained by the automatic annotation process by a physician or another computer (other software). The annotation device 1660 used for verification by a physician may have a configuration similar to that of the annotation device 1660 used for the annotation process. Similarly, the annotation device 1660 used for verification by another computer (other software) may have a configuration similar to that of the annotation device 1660 used for the annotation process.

[0134] As described above, pairs of data and annotation information are formed by the data source 1640, the data acquisition device 1650, and the annotation device 1660. The training data 1670 contains multiple pairs of data and annotation information. In other words, the training data 1670 contains a collection of pairs of data and annotation information.

[0135] The data included in the training data 1670 is not limited to data generated using the data source 1640, the data acquisition device 1650, and the annotation device 1660 (data acquired by the data acquisition device 1650, annotation information acquired using the annotation device 1660). For example, the training data 1670 may include any of the following types of data: data acquired by devices other than the data acquisition device 1650; data generated by processing data acquired by the data acquisition device 1650 or other devices; computer-generated data; pseudo-data; data generated by data augmentation. The training data 1670 may also include annotation information associated with this data.

[0136] The learning processing unit 1610 can construct the neural network 1221 by applying supervised learning (and / or semi-supervised learning) using the training data 1670 to the neural network 1620. This supervised learning is performed, for example, so that for each pair (a pair of data and annotation information) included in the training data 1670, the annotation information is obtained as the output corresponding to the input of the data.

[0137] As mentioned above, the data in each pair included in the training data 1670 of this example (i.e., the data associated with the annotation information) includes data of the same type as the reference data (first type of data) and data of the same type as the target data (second type of data). Furthermore, the annotation information in each pair included in the training data 1670 of this example (i.e., the annotation information associated with the data) includes information of the same type as the consistency evaluation information.

[0138] Therefore, the inference model 1220, which includes the neural network 1221 in this example, functions to receive pairs of data of a first type and data of a second type as input and output consistency evaluation information. In other words, the inference model 1220, which includes the neural network 1221 in this example, is configured to receive pairs of reference data and target data as input and output consistency evaluation information.

[0139] When the evaluation unit 1200 is capable of performing consistency evaluation using a machine learning model, machine learning to create the machine learning model can be performed for each user (healthcare professional). In other words, the consistency evaluation of medical data generated by a specific user can be performed using a machine learning model corresponding to that user.

[0140] In this example, the training data includes a second type of data (the same type as the target data), which is data previously entered by a specific healthcare professional. Furthermore, the machine learning model constructed using this training data receives newly entered data from this healthcare professional as the target data and outputs consistency evaluation information.

[0141] As one specific example, the training data includes a set of electronic medical record data previously created by a particular physician. By performing machine learning using this training data, a machine learning model that has learned the habits and tendencies of this physician is obtained. By inputting newly created electronic medical record data by this physician into the evaluation unit 1200 equipped with this machine learning model, consistency evaluation information that reflects the habits and tendencies of this physician can be obtained. At the very least, it is considered that consistency evaluation information that reflects the habits and tendencies of the physician can be obtained more accurately than when using a machine learning model created using machine learning with a set of electronic medical record data created by multiple physicians.

[0142] In some exemplary embodiments, the evaluation unit 1200 may be configured to perform consistency evaluation using a machine learning model that has learned clinical practice guidelines for a specific disease.

[0143] In this example, the training data includes standard data containing standardized information on medical procedures as the first type of data (data of the same type as the reference data), and furthermore, this standard data includes clinical practice guidelines for a specific disease. A machine learning model (the second machine learning model) constructed using machine learning with this training data receives target data related to this disease as input and outputs consistency evaluation information.

[0144] By inputting target data (e.g., electronic medical record data) into the evaluation unit 1200 equipped with the machine learning model in this example, it is possible to evaluate whether the content of the input target data conforms to the clinical practice guidelines for the disease in question. In other words, it is possible to evaluate the consistency between the content of the input target data and the clinical practice guidelines for the disease in question. This makes it possible to detect, for example, that tests recommended by the clinical practice guidelines are missing, or that decisions have been made that do not conform to the criteria shown in the clinical practice guidelines. Furthermore, by combining this with the real-time processing described above, it is possible to provide information and warnings based on the clinical practice guidelines.

[0145] The assessment of the consistency of medical data based on clinical practice guidelines is not limited to methods using machine learning models as described above; it is also possible to implement consistency assessment without using machine learning models. In some of the examples described below, machine learning models may be used, at least partially, or not at all.

[0146] In some exemplary embodiments, the reference data may include clinical practice guidelines for a specific disease as standard data. Furthermore, the evaluation unit 1200 may be configured to receive input of target data and perform consistency evaluation by determining whether information regarding predetermined matters described in the clinical practice guidelines is included in the target data.

[0147] The "specified items" described above in the clinical practice guidelines are predetermined and / or set for each consistency evaluation.

[0148] If the prescribed items in the clinical practice guidelines include questions to be asked during the interview, the evaluation unit 1200 can perform a consistency evaluation by determining whether the interview data (interview results) corresponding to these questions are included in the target data (for example, target input data such as electronic medical record data entered by a physician). This makes it possible to determine whether the interview was conducted in accordance with the clinical practice guidelines and whether the results of the interview, which were conducted in accordance with the clinical practice guidelines, have been recorded.

[0149] If the clinical practice guidelines include a specified item for the type of test (the type of test to be performed), the evaluation unit 1200 can perform a consistency evaluation by determining whether the test data (test results) corresponding to this test type is included in the target data (for example, the target input data such as electronic medical record data entered by a physician). This makes it possible to determine whether the test was performed in accordance with the clinical practice guidelines and whether the results of the test, which were performed in accordance with the clinical practice guidelines, have been recorded.

[0150] The annotation unit 1300 is configured to add annotation information to the reference data and / or target data based on the results of the consistency evaluation between medical data performed by the evaluation unit 1200.

[0151] The annotation information assigned to the consistency evaluation results by the annotation unit 1300 is different from the annotation information assigned to the data by the annotation device 1660 in order to create the training data 1670. This information is generated based on the consistency evaluation results obtained by the evaluation unit 1200 and is assigned to the data (reference data, target data) input to the information evaluation unit 1200.

[0152] In some exemplary embodiments, the annotation unit 1300 may be configured to generate and add annotation information when it is determined that there is no consistency (or doubt about consistency) between the reference data and the target data input to the evaluation unit 1200. Examples of information included in this annotation information include warnings, notifications, suggestions for additions to improve consistency and / or suggestions for additions, and suggestions for corrections to improve consistency and / or suggestions for corrections.

[0153] In some exemplary embodiments, the annotation unit 1300 may be configured to generate and add annotation information even when it is determined that there is consistency (no doubt about consistency) between the reference data and the target data input to the evaluation unit 1200. Examples of information to be included in the annotation information in this case include messages or icons indicating that there are no problems with the consistency of the medical data.

[0154] Such annotation information is typically attached to the target data, but it may also be attached to reference data that has been determined to be inconsistent with a large number of target data. In other words, this annotation information can be said to raise questions about reference data that is inconsistent with a large number of data collected in clinical settings.

[0155] The functions of the annotation unit 1300 are realized through the cooperation of software, such as an annotation program, and hardware, such as a processor. The annotation unit 1300 operates under the control of a processor (control unit), for example, not shown. This processor is configured to control the annotation unit 1300 by a control program, for example, not shown.

[0156] The output unit 1400 has a configuration for outputting data from the medical data processing system 1000. In this embodiment, the output unit 1400 includes a display control unit 1410. The display control unit 1410 performs control to display information on the display device 1510 of the user interface 1500. The functions of the output unit 1400 (display control unit 1410) are realized through the cooperation of software such as an output program (display control program) and hardware such as a processor.

[0157] The user interface 1500 includes a display device 1510 and an operating device 1520. The operating device 1520 includes various operating devices and input devices. The user interface 1500 may also include a device that integrates display and operating functions, such as a touch panel. At least a part of the user interface 1500 may be located outside the medical data processing system 1000. For example, the display device 1510 and the operating device 1520 may be peripheral devices of the medical data processing system 1000.

[0158] Here are some examples of real-time processing.

[0159] In several exemplary embodiments, the medical data processing system 1000 can perform consistency evaluations in parallel with the data entry work being performed by the user. In this example, the target data includes new data entered by a healthcare professional during their current work as target input data. The evaluation unit 1200 can then perform a consistency evaluation between the reference data and this new data in response to receiving this new data.

[0160] As a specific example, the target data may include new data entered by a physician during the electronic medical record entry process currently being performed. Furthermore, the evaluation unit 1200 can perform a consistency evaluation between the reference data and this new data in response to receiving this new data. The medical data processing system 1000 may be configured to perform such a consistency evaluation each time a physician enters new data into the electronic medical record.

[0161] This real-time consistency evaluation makes it possible to evaluate the accuracy of newly entered data in real time during tasks currently being performed by healthcare professionals. Furthermore, by performing annotation information generation by the annotation unit 1300 in real time, it becomes possible to evaluate the accuracy of newly entered data in real time during tasks currently being performed by healthcare professionals and to provide annotation information based on the results of this evaluation in real time.

[0162] In addition to such real-time consistency evaluation (and real-time annotation information generation), the medical data processing system 1000 may be configured to output the consistency evaluation results obtained from real-time processing in real time.

[0163] In this example, the evaluation unit 1200 performs a real-time consistency evaluation between new data entered by the healthcare professional during the work currently being performed and the reference data, and the output unit 1400 outputs the consistency evaluation results obtained from this real-time processing in real time.

[0164] This combination of real-time consistency evaluation and real-time output makes it possible to evaluate the accuracy of newly entered data in real time during tasks currently being performed by healthcare professionals, and to provide the results of this evaluation to those healthcare professionals in real time. Furthermore, by combining this with real-time annotation information generation, it becomes possible to evaluate the accuracy of newly entered data in real time during tasks currently being performed by healthcare professionals, and to provide the results of this evaluation and the resulting annotation information to those healthcare professionals in real time.

[0165] The method of outputting the consistency evaluation results in real time is arbitrary. In the medical data processing system 1000 with the configuration shown in Figure 2, the display control unit 1410 can display the consistency evaluation results on the display device 1510 in real time.

[0166] As another example, the output unit 1400A of the medical data processing system 1000A shown in Figure 6 includes a transmission unit 1420. The transmission unit 1420 includes communication devices such as a modem, router, and communication circuit.

[0167] In this example, the evaluation unit 1200 performs a real-time consistency evaluation between new data entered by the healthcare worker during their current work and reference data, and the transmission unit 1420 transmits the consistency evaluation results (and annotation information) obtained from this real-time processing to the computer 2700 used by the healthcare worker in real time. The computer 2700 is, for example, the healthcare worker's terminal 2600.

[0168] Computer 2700 receives information transmitted from the medical data processing system 1000A via the communication unit 2710. The received information is stored in the storage unit 2720 and displayed on the display unit 2730.

[0169] This combination of real-time consistency evaluation and real-time transmission makes it possible to evaluate the accuracy of newly entered data in real time during the work currently being performed by healthcare professionals, and to provide the results of this evaluation to the healthcare professionals in real time. Furthermore, by combining this with real-time annotation information generation, it becomes possible to evaluate the accuracy of newly entered data in real time during the work currently being performed by healthcare professionals, and to provide the results of this evaluation and the annotation information based on it to the healthcare professionals in real time. In particular, one of the features of this example is that the same effect can be obtained even when healthcare professionals are working in a location far from the medical data processing system 1000A (for example, a remote location).

[0170] Several examples of the operation of the medical data processing system 1000 according to this embodiment will be described. Two or more operation examples can be arbitrarily selected from these examples and combined at least partially. Furthermore, the operation of the medical data processing system 1000 is not limited to these operation examples, and for example, any matters described in this disclosure can be combined with these operation examples.

[0171] The first example of operation will be explained with reference to Figure 7. This example of operation provides one aspect of the medical data processing method according to the embodiment.

[0172] First, the reference data and target data are prepared (S1, S2). The reference data includes one or more of the following: reference input data, reference test data, and standard data. The target data includes target input data and / or target test data. The timing for acquiring the reference data and the timing for acquiring the target data are both arbitrary. Steps S1 and S2 are performed by the medical data acquisition unit 1100.

[0173] Next, an evaluation of the consistency between the reference data acquired in step S1 and the target data acquired in step S2 is performed (S3). Step S3 is performed by the evaluation unit 1200.

[0174] According to this example, it is possible to improve the overall quality of medical data (for example, all medical data concerning a particular patient) by evaluating the consistency between reference data and target data.

[0175] A second example of operation will be described with reference to Figure 8. This example of operation provides one aspect of the medical data processing method according to the embodiment.

[0176] Similar to the first example of operation, the medical data acquisition unit 1100 prepares the reference data and the target data (S11, S12), and the evaluation unit 1200 performs an evaluation of the consistency between the reference data acquired in step S11 and the target data acquired in step S12 (S13).

[0177] Furthermore, based on the results of the consistency evaluation performed in step S13, annotation information is added to the reference data acquired in step S11 and / or the target data acquired in step S12 (S14). Step S14 is performed by the annotation unit 1300.

[0178] According to this example, by evaluating the consistency between reference data and target data, it is possible to improve the overall quality of medical data (for example, the entire medical data concerning a particular patient). In addition, information (annotation information) corresponding to the results of the consistency evaluation can be attached to the reference data and / or target data. For example, if there is no consistency between the reference data and target data, annotation information including warnings and suggestions can be attached to the reference data and / or target data. If there is consistency between the reference data and target data, annotation information including messages and icons indicating that there are no problems with the consistency of the medical data can be attached to the reference data and / or target data.

[0179] A third example of operation will be described with reference to Figure 9. This example of operation provides one aspect of the medical data processing method according to the embodiment, and provides one aspect of the real-time processing described above.

[0180] First, a healthcare professional begins data entry using the healthcare professional terminal 2600 (S21). For example, a doctor begins electronic medical record entry. Various information is referenced during the data entry process. For example, test data and image interpretation reports are referenced during electronic medical record entry.

[0181] In this example, as in the first example, the medical data acquisition unit 1100 prepares the reference data (S22). The timing of preparing the reference data can be arbitrary.

[0182] Reference data may, but is not limited to, data referenced in data entry work. For example, if a disease name is included in the electronic medical record data being entered, the clinical practice guidelines for this disease can be prepared as reference data. This preparation process is performed, for example, by the following series of steps: (1) The healthcare worker terminal 2600 searches for the disease name in the electronic medical record data; (2) The medical data acquisition unit 1100 transmits the searched disease name to the medical data processing system 1000; (3) The medical data acquisition unit 1100 of the medical data processing system 1000 acquires the clinical practice guidelines corresponding to this disease name.

[0183] Here, the storage location for the clinical practice guidelines can be arbitrary and may be, for example, the medical data processing system 1000, the hospital information system 2100, or the departmental system 2400. Alternatively, if the clinical practice guidelines are stored in the medical professional terminal 2600, the medical professional terminal 2600 may search for disease names in the electronic medical record data, retrieve the clinical practice guidelines corresponding to the retrieved disease names, and transmit the retrieved clinical practice guidelines to the medical data processing system 1000.

[0184] When the medical professional who started the data entry work in step S21 enters new data into the medical professional terminal 2600 (S23), this new input data is sent to the medical data processing system 1000. The medical data acquisition unit 1100 of the medical data processing system 1000 accepts this new input data as target data (S24).

[0185] The types of input data adopted as target data may be arbitrary, but including all data entered by medical professionals as target data may unnecessarily increase the processing load of the medical data processing system 1000. Taking this into consideration, the medical professional terminal 2600 (or a relay device operating between the medical professional terminal 2600 and the medical data processing system 1000) may be configured to perform a process of selecting the data to be provided to the medical data processing system 1000.

[0186] For example, the medical professional terminal 2600 may be configured to hold a pre-created list of strings (a list of words or sentences, etc.), and when a string included in this list or a similar string is entered, it may be sent to the medical data processing system 1000 as new input data. The string list may be provided for each medical department, each disease, each disease group, or each medical professional.

[0187] As another example, the healthcare worker terminal 2600 may have a machine learning model that functions to select data from the data entered by the healthcare worker to send to the medical data processing system 1000.

[0188] The evaluation unit 1200 performs an evaluation of the consistency between the reference data acquired in step S22 and the target data acquired in step S24 (S25).

[0189] This example demonstrates that by evaluating the consistency between reference data and target data, it is possible to improve the overall quality of medical data (for example, the entire medical data for a particular patient), and also to provide healthcare professionals performing data entry tasks with real-time warnings and suggestions regarding consistency.

[0190] A fourth example of operation will be described with reference to Figure 10. This example of operation provides one aspect of the medical data processing method according to the embodiment, and provides one aspect of the process for evaluating the consistency of test data.

[0191] First, a first medical examination is performed on the patient using the examination device 2500, and first data is acquired (S31). This first data is sent from the examination device 2500 to the medical data processing system 1000. The examination data acquisition unit 1110 of the medical data processing system 1000 generates first examination data from the first data acquired in step S31 (S32), and further, the medical data acquisition unit 1100 generates reference data including this first examination data (S33).

[0192] Similarly, a second medical examination is performed on the patient using the examination device 2500, and second data is acquired (S34). This second data is sent from the examination device 2500 to the medical data processing system 1000. The examination data acquisition unit 1110 of the medical data processing system 1000 generates second examination data from the second data acquired in step S34 (S35), and further, the medical data acquisition unit 1100 generates target data including this second examination data (S36).

[0193] The evaluation unit 1200 performs an evaluation of the consistency between the reference data acquired in step S33 and the target data acquired in step S36 (S37).

[0194] According to this example, the overall quality of medical data (for example, the entire medical data concerning a particular patient) can be improved by evaluating the consistency between reference data, which includes test data, and target data, which includes other test data.

[0195] The fifth example of operation will be explained with reference to Figure 11. This example of operation provides one aspect of the medical data processing method according to the embodiment, and provides one aspect of the process of evaluating consistency using a machine learning model.

[0196] First, the data acquisition device 1650 collects data for creating training data from the data source 1640 (S41). The type of data collected in step S41 is determined based on the design of the machine learning model constructed in this example. In step S41, a first type of data is collected, which is the same type as the reference data prepared in step S46 described later, and a second type of data is collected, which is the same type as the target data prepared in step S47 described later.

[0197] The data collected in step S41 is sent to the annotation device 1660. The annotation device 1660 performs the process of adding annotation information to the collected data (S42). Based on the data collected in step S41 and the annotation information added in step S42, training data 1670 is created (S43). The created training data 1670 is provided to the inference model building device 1600.

[0198] The learning processing unit 1610 of the inference model construction device 1600 constructs a machine learning model by applying machine learning using the training data 1670 created in step S43 to the neural network 1620 (S44). The constructed machine learning model is provided to the evaluation unit 1200 of the medical data processing system 1000 (S45).

[0199] Similar to the first example of operation, the medical data acquisition unit 1100 prepares the reference data and the target data (S46, S47). The evaluation unit 1200 inputs the reference data acquired in step S46 and the target data acquired in step S47 into the machine learning model constructed in step S44 (S48). The machine learning model generates consistency evaluation information in response to the input of the reference data and the target data (S49). The evaluation unit 1200 may create a consistency evaluation result based on the consistency evaluation information generated in step S49.

[0200] This example demonstrates that a machine learning model can be used to evaluate the consistency between reference data and target data, thereby enabling further improvement in the overall quality of medical data (for example, the entire set of medical data concerning a particular patient).

[0201] The sixth example of operation will be explained with reference to Figure 12. This example of operation provides one aspect of the medical data processing method according to the embodiment, and provides one aspect of the process of evaluating the consistency of medical data with respect to clinical practice guidelines using a machine learning model.

[0202] First, the data acquisition device 1650 obtains clinical practice guidelines for a specific disease from the data source 1640 (S51). The type of clinical practice guideline obtained in step S51 is determined based on the design of the machine learning model constructed in this example.

[0203] Based on the clinical guidelines obtained in step S51, training data 1670 is created (S52). The created training data 1670 is provided to the inference model building device 1600.

[0204] Furthermore, if annotation is to be performed based on the clinical practice guideline obtained in step S51, this clinical practice guideline is sent to the annotation device 1660. In this case, training data 1670 is created based on the clinical practice guideline obtained in step S51 and the annotation information generated by the annotation device 1660 (S52).

[0205] The learning processing unit 1610 of the inference model construction device 1600 constructs a machine learning model by applying machine learning using the training data 1670 created in step S52 to the neural network 1620 (S53). The constructed machine learning model is provided to the evaluation unit 1200 of the medical data processing system 1000 (S54).

[0206] Similar to the first example of operation, the medical data acquisition unit 1100 prepares the target data (S55). The evaluation unit 1200 inputs the target data acquired in step S55 into the machine learning model constructed in step S53 (S56). The machine learning model generates consistency evaluation information in response to the input of the target data (S57). The evaluation unit 1200 may create a consistency evaluation result based on the consistency evaluation information generated in step S57.

[0207] This example demonstrates that a machine learning model can be used to evaluate the consistency of target data with respect to clinical practice guidelines (reference data), thereby enabling further improvement in the overall quality of medical data (for example, all medical data related to a particular patient).

[0208] The seventh example of operation will be explained with reference to Figure 13. This example of operation provides one aspect of the medical data processing method according to the embodiment, and provides one aspect of the process for evaluating the consistency of medical data with respect to clinical practice guidelines.

[0209] First, the medical data acquisition unit 1100 acquires clinical practice guidelines for a specific disease (S61) and prepares reference data including these clinical practice guidelines (S62). Also, similar to the first example of operation, the medical data acquisition unit 1100 prepares the target data (S63).

[0210] The evaluation unit 1200 determines whether the prescribed items included in the clinical practice guidelines in the reference data created in step S62 are included in the target data acquired in step S63 (S64). Furthermore, based on the result of the determination in step S64, the evaluation unit 1200 performs an evaluation of the consistency between the clinical practice guidelines in the reference data created in step S62 and the target data acquired in step S63 (S65).

[0211] This example demonstrates that it is possible to evaluate whether the target data conforms to clinical guidelines (reference data), thereby improving the overall quality of medical data (for example, the entire medical data related to a particular patient).

[0212] The eighth example of operation will be explained with reference to Figure 14. This example of operation provides one aspect of the medical data processing method according to the embodiment, and provides one aspect of the process of evaluating the consistency of medical data with respect to clinical practice guidelines using a machine learning model.

[0213] First, similar to the sixth example of operation, the data acquisition device 1650 obtains clinical practice guidelines for a specific disease from the data source 1640 (S71). Furthermore, similar to the fifth example of operation, the data acquisition device 1650 collects data for creating training data from the data source 1640 (S72). The data collected in step S72 may be the same type of data as the target data prepared in step S76 described later.

[0214] Training data 1670 is created based on the clinical guidelines obtained in step S71 and the data collected in step S72 (S73). The created training data 1670 is provided to the inference model building device 1600.

[0215] When annotation is performed based on the clinical guidelines obtained in step S71 and / or the data obtained in step S72, the clinical guidelines and / or the data are sent to the annotation device 1660. In this case, training data 1670 is created based on the clinical guidelines obtained in step S71, the data collected in step S72, and the annotation information generated by the annotation device 1660 (S73).

[0216] The learning processing unit 1610 of the inference model construction device 1600 constructs a machine learning model (S74) by applying machine learning using the training data 1670 created in step S73 to the neural network 1620. This machine learning model has the function of comparing (comparing) medical data with clinical practice guidelines. The machine learning model constructed in step S74 is provided to the evaluation unit 1200 of the medical data processing system 1000 (S75).

[0217] Similar to the first example of operation, the medical data acquisition unit 1100 prepares the target data (S76). The evaluation unit 1200 inputs the target data acquired in step S76 into the machine learning model constructed in step S74 (S77). The machine learning model generates consistency evaluation information in response to the input of the target data (S78). This consistency evaluation information includes information (comparison information) showing the results of comparing the target data acquired in step S76 with the clinical practice guidelines acquired in step S71. The evaluation unit 1200 creates a consistency evaluation result based on the consistency evaluation information generated in step S78 (S79).

[0218] This example demonstrates that it is possible to use a machine learning model to evaluate whether the target data conforms to clinical guidelines (reference data), thereby enabling further improvement in the overall quality of medical data (for example, the entire medical data related to a particular patient).

[0219] <Medical data processing device> Several exemplary embodiments of the medical data processing device according to the embodiment will be described.

[0220] In some exemplary embodiments, the medical data processing device may have a configuration similar to that of the medical data processing system 1000 in Figure 2 or the medical data processing system 1000A in Figure 6. The medical data processing device may be an element of an information system within a medical institution, as in the medical data processing system 1000 in Figure 1, but it may also be located on a wide-area network such as the Internet. A medical data processing device located on a wide-area network may be configured, for example, to provide cloud services to multiple information systems within medical institutions.

[0221] The operation, function, and effects of the medical data processing device according to the embodiment are the same as those of the medical data processing system according to the embodiment; therefore, please refer to the descriptions of the various exemplary embodiments of the medical data processing system described above. Furthermore, any matters relating to the medical data processing system can be combined with the medical data processing device.

[0222] The medical data processing device according to this embodiment includes a receiving unit that receives first medical data (reference data) and second medical data (target data). This receiving unit may be the same as the medical data acquisition unit 1100 in Figure 2. In a medical data processing device that provides cloud services to multiple in-house information systems of medical institutions, the receiving unit is equipped with a communication device for data communication with each in-house information system of a medical institution via a wide-area network.

[0223] The evaluation unit of the medical data processing device is configured to perform an evaluation of the consistency between the first medical data and the second medical data, similar to the evaluation unit 1200 in Figure 2. Here, the first medical data includes one of the following: first input data (reference input data) entered by a medical professional, first test data (reference test data) obtained based on data obtained by a medical test, and pre-created standard data. The second medical data includes second input data (target input data) entered by a medical professional, and / or second test data (target test data) obtained based on data obtained by a medical test (see Figure 3).

[0224] In some exemplary embodiments, the medical data processing device may further include an annotation unit that assigns annotation information to the first medical data and / or the second medical data based on the results of a consistency evaluation performed by the evaluation unit. This annotation unit may be similar to the annotation unit 1300 in Figure 2.

[0225] In some exemplary embodiments, the receiving unit of the medical data processing device can receive new data entered by a healthcare professional during work currently being performed as second input data for the second medical data. Furthermore, the evaluation unit may be configured to perform an evaluation of the consistency between the first medical data and the new data in response to the receipt of the new data by the receiving unit.

[0226] In some exemplary embodiments, the receiving unit of the medical data processing device can receive data obtained by medical examinations. Furthermore, the medical data processing device may further include an examination data acquisition unit that acquires examination data based on the data received by the receiving unit. This examination data acquisition unit may be similar to the examination data acquisition unit 1110 in Figure 2. In this embodiment, if the first medical data includes first examination data, this first examination data may be acquired by the examination data acquisition unit, and / or, if the second medical data includes second examination data, this second examination data may be acquired by the examination data acquisition unit.

[0227] In some exemplary embodiments, the evaluation unit of the medical data processing device may include a first machine learning model. The first machine learning model is constructed by machine learning using training data that includes a first type of data identical to the first medical data and a second type of data identical to the second medical data. Furthermore, the first machine learning model functions to receive input from at least the first medical data and the second medical data and output consistency evaluation information.

[0228] In some exemplary embodiments, the first medical data may include clinical practice guidelines for a specific disease as standard data. Furthermore, the evaluation unit of the medical data processing device may include a second machine learning model. The second machine learning model is constructed by machine learning using training data including these clinical practice guidelines, and functions to receive at least the second medical data as input and output consistency evaluation information.

[0229] In some exemplary embodiments, the first medical data may include a clinical practice guideline for a specific disease as standard data. Furthermore, the evaluation unit of the medical data processing device may be configured to evaluate the consistency of the second medical data with respect to the clinical practice guideline by determining whether the second medical data contains information regarding predetermined matters described in the clinical practice guideline. The evaluation unit in this embodiment may be configured to perform consistency evaluation using a machine learning model at least in part, or it may be configured to perform consistency evaluation without using a machine learning model.

[0230] <Medical Data Processing Methods> According to the embodiment of the medical data processing system, its exemplary embodiments, the embodiment of the medical data processing device, its exemplary embodiments, etc., it is possible to realize various medical data processing methods. Several exemplary embodiments are described below.

[0231] The steps, operation, and effects of the medical data processing method according to the embodiment are the same as those of the medical data processing system according to the embodiment; therefore, please refer to the descriptions of the various exemplary embodiments of the medical data processing system described above. Furthermore, any matters relating to the medical data processing system can be combined with the medical data processing method.

[0232] The medical data processing method according to this embodiment is a method for processing medical data to support the medical treatment of a patient, and includes the following steps: (1) a step in which a computer prepares first medical data including either first input data entered by a medical professional, first test data obtained based on data obtained by a medical examination, or pre-created standard data; (2) a step in which a computer prepares second medical data including either second input data entered by a medical professional or second test data obtained based on data obtained by a medical examination; (3) a step in which a computer performs an evaluation of the consistency between the first medical data and the second medical data. The example of operation of the medical data processing system 1000 shown in Figure 7 provides one example of this medical data processing method.

[0233] In some exemplary embodiments, the medical data processing method may further include a step in which a computer adds annotation information to first medical data and / or second medical data based on the results of a integrity assessment. The example of operation of the medical data processing system 1000 shown in Figure 8 provides one example of this medical data processing method.

[0234] In some exemplary embodiments, the step of preparing second medical data may include a step in which the computer receives new data entered by a healthcare professional in the course of work currently being performed, as second input data. Furthermore, the step of performing consistency evaluation may include a step in which the computer performs consistency evaluation between the first medical data and this new data in response to the receipt of this new data. The example of operation of the medical data processing system 1000 shown in Figure 9 provides one example of this medical data processing method.

[0235] In some exemplary embodiments, the steps of preparing first test data and / or second test data may include the steps of a computer preparing data obtained by a medical examination and the steps of a computer obtaining test data based on this data obtained by the medical examination. The example of operation of the medical data processing system 1000 shown in Figure 10 provides one example of this medical data processing method.

[0236] In some exemplary embodiments, the process of performing consistency evaluation may be carried out using a first machine learning model. This first machine learning model is constructed by machine learning using training data that includes a first type of data identical to the first medical data and a second type of data identical to the second medical data, and functions to receive at least the first and second medical data as inputs and output consistency evaluation information. The example of operation of the medical data processing system 1000 shown in Figure 11 provides one example of this medical data processing method.

[0237] In some exemplary embodiments, the first medical data may include clinical practice guidelines for a specific disease as standard data. Furthermore, the process of performing consistency assessment may be carried out using a second machine learning model. This second machine learning model is constructed by machine learning using training data including these clinical practice guidelines and functions to receive at least the second medical data as input and output consistency assessment information. The example of operation of the medical data processing system 1000 shown in Figure 12 provides one example of this medical data processing method.

[0238] In some exemplary embodiments, the first medical data may include, as standard data, clinical practice guidelines for a particular disease. Furthermore, the step of performing a consistency assessment may include a step in which a computer determines whether the second medical data contains information about predetermined matters described in these clinical practice guidelines. The operation examples of the medical data processing system 1000 shown in Figures 13 and 14 provide two examples of this medical data processing method.

[0239] <Programs and recording media> It is possible to construct a program that causes a computer to perform any one or more of the processes (steps) described in this disclosure. It is also possible to create a recording medium on which such a program is stored. The recording medium is a non-temporary recording medium that can be read by a computer. The form of such a recording medium is arbitrary, and examples include magnetic disks, optical disks, magneto-optical disks, and semiconductor memory.

[0240] This disclosure presents several exemplary embodiments of the invention. These embodiments are merely illustrative of the present invention. Therefore, any modifications (omissions, substitutions, additions, etc.) within the scope of the gist of the invention can be applied to this disclosure. [Explanation of symbols]

[0241] 1000 Medical Data Processing Systems 1100 Medical Data Acquisition Department 1110 Inspection Data Acquisition Unit 1200, 1200A Evaluation Unit 1210 Judgment section 1220 Inference Models 1300 Annotation Section 1400, 1400A output section 1410 Display Control Unit 1420 Transmitter 1500 User Interfaces 1510 Display device 1520 Operating device

Claims

1. A system for processing medical data, Includes an evaluation unit that performs an evaluation of the consistency between the first medical data and the second medical data, The first medical data includes any of the following: first input data entered by a medical professional, first test data obtained based on data obtained by a medical examination, and pre-created standard data. The second medical data includes either second input data entered by a healthcare professional, or second test data obtained based on data obtained through medical tests. The first medical data includes a first data group consisting of two or more data, and / or the second medical data includes a second data group consisting of two or more data. The evaluation unit performs the evaluation of consistency based on the first data group and / or the second data group. The first data set includes standard range information for the examination of a predetermined item and relational information indicating the relationship between the examination of the predetermined item and a predetermined medical procedure. The second data set includes test data obtained by applying the predetermined tests to a patient and medical procedure information indicating the medical procedures performed on the patient. The evaluation unit, A determination is made, based on the relationship information, whether or not there is a relationship between the examination of the predetermined items and the medical procedure indicated in the medical procedure information. If it is determined that there is no such relationship, the standard range information and the inspection data are compared, and the consistency evaluation is performed based on the results of the comparison. If the aforementioned relationship is determined to exist, the standard range information and the inspection data are compared, processing is applied to the comparison result based on the relationship information, and the consistency evaluation is performed based on the result of the processing. system.

2. The aforementioned prescribed medical procedure includes prescribed surgery, The aforementioned medical procedure information includes surgical information indicating the surgery performed on the patient, The evaluation unit, as the determination based on the relational information, performs a determination as to whether or not there is a relationship between the examination of the predetermined item and the surgery indicated in the surgical information. The system according to claim 1.

3. The aforementioned prescribed medical procedure includes the administration of a prescribed drug, The aforementioned medical information includes drug information indicating the medication prescribed to the patient. The evaluation unit, as the determination based on the relationship information, performs a determination as to whether or not there is a relationship between the inspection of the predetermined item and the drug indicated in the drug information. The system according to claim 1.

4. A system for processing medical data, Includes an evaluation unit that performs an evaluation of the consistency between the first medical data and the second medical data, The first medical data includes any of the following: first input data entered by a medical professional, first test data obtained based on data obtained by a medical examination, and pre-created standard data. The second medical data includes either second input data entered by a healthcare professional, or second test data obtained based on data obtained through medical tests. The first medical data includes, as the first input data, information on the treatment site of the patient. The second medical data includes, as the second examination data, treatment image data acquired during the treatment of the patient. The evaluation unit identifies the treatment target site based on the treatment image data and performs the consistency evaluation by comparing the identified target site with the treatment target site information. system.

5. A system for processing medical data, Includes an evaluation unit that performs an evaluation of the consistency between the first medical data and the second medical data, The first medical data includes any of the following: first input data entered by a medical professional, first test data obtained based on data obtained by a medical examination, and pre-created standard data. The second medical data includes either second input data entered by a healthcare professional, or second test data obtained based on data obtained through medical tests. The aforementioned medical examinations include ophthalmological examinations, The examination data obtained by the aforementioned ophthalmic examination includes at least one of the following: intraocular pressure data, optical coherence tomography data, captured image data, strabismus data, heterophoria data, visual acuity data, depth perception data, refractive power data, ocular aberration data, visual field data, corneal morphology data, fundus morphology data, corneal endothelial cell data, axial length data, binocular vision function data, and color vision data. system.

6. A device for processing medical data, A reception area that receives the first medical data and the second medical data, An evaluation unit that performs an evaluation of the consistency between the first medical data and the second medical data. Includes, The first medical data includes any of the following: first input data entered by a medical professional, first test data obtained based on data obtained by a medical examination, and pre-created standard data. The second medical data includes either second input data entered by a healthcare professional, or second test data obtained based on data obtained through medical tests. The first medical data includes a first data group consisting of two or more data, and / or the second medical data includes a second data group consisting of two or more data. The evaluation unit performs the evaluation of consistency based on the first data group and / or the second data group. The first data set includes standard range information for the examination of a predetermined item and relational information indicating the relationship between the examination of the predetermined item and a predetermined medical procedure. The second data set includes test data obtained by applying the predetermined tests to a patient and medical procedure information indicating the medical procedures performed on the patient. The evaluation unit, A determination is made, based on the relationship information, whether or not there is a relationship between the examination of the predetermined items and the medical procedure indicated in the medical procedure information. If it is determined that there is no such relationship, the standard range information and the inspection data are compared, and the consistency evaluation is performed based on the results of the comparison. If the aforementioned relationship is determined to exist, the standard range information and the inspection data are compared, processing is applied to the comparison result based on the relationship information, and the consistency evaluation is performed based on the result of the processing. Device.

7. A device for processing medical data, A reception area that receives the first medical data and the second medical data, An evaluation unit that performs an evaluation of the consistency between the first medical data and the second medical data. Includes, The first medical data includes any of the following: first input data entered by a medical professional, first test data obtained based on data obtained by a medical examination, and pre-created standard data. The second medical data includes either second input data entered by a healthcare professional, or second test data obtained based on data obtained through medical tests. The first medical data includes, as the first input data, information on the treatment site of the patient. The second medical data includes, as the second examination data, treatment image data acquired during the treatment of the patient. The evaluation unit identifies the treatment target site based on the treatment image data and performs the consistency evaluation by comparing the identified target site with the treatment target site information. Device.

8. A device for processing medical data, A reception area that receives the first medical data and the second medical data, An evaluation unit that performs an evaluation of the consistency between the first medical data and the second medical data. Includes, The first medical data includes any of the following: first input data entered by a medical professional, first test data obtained based on data obtained by a medical examination, and pre-created standard data. The second medical data includes either second input data entered by a healthcare professional, or second test data obtained based on data obtained through medical tests. The aforementioned medical examinations include ophthalmological examinations, The examination data obtained by the aforementioned ophthalmic examination includes at least one of the following: intraocular pressure data, optical coherence tomography data, captured image data, strabismus data, heterophoria data, visual acuity data, depth perception data, refractive power data, ocular aberration data, visual field data, corneal morphology data, fundus morphology data, corneal endothelial cell data, axial length data, binocular vision function data, and color vision data. Device.

9. A method for processing medical data to support patient care, Prepare a first medical data set that includes either the first input data entered by a healthcare professional, the first test data obtained based on data acquired through medical tests, or pre-created standard data. Prepare second medical data, which includes either second input data entered by a healthcare professional, or second test data obtained based on data obtained through medical tests. An evaluation of the consistency between the first medical data and the second medical data is performed. The first medical data includes a first data group consisting of two or more data, and / or the second medical data includes a second data group consisting of two or more data. Based on the first data set and / or the second data set, the consistency evaluation is performed. The first data set includes standard range information for the examination of a predetermined item and relational information indicating the relationship between the examination of the predetermined item and a predetermined medical procedure. The second data set includes test data obtained by applying the predetermined tests to a patient and medical procedure information indicating the medical procedures performed on the patient. A determination is made, based on the relationship information, whether or not there is a relationship between the examination of the predetermined items and the medical procedure indicated in the medical procedure information. If it is determined that there is no such relationship, the standard range information and the inspection data are compared, and the consistency evaluation is performed based on the results of the comparison. If the aforementioned relationship is determined to exist, the standard range information and the inspection data are compared, processing is applied to the comparison result based on the relationship information, and the consistency evaluation is performed based on the result of the processing. method.

10. A method for processing medical data to support patient care, Prepare a first medical data set that includes either the first input data entered by a healthcare professional, the first test data obtained based on data acquired through medical tests, or pre-created standard data. Prepare second medical data, which includes either second input data entered by a healthcare professional, or second test data obtained based on data obtained through medical tests. An evaluation of the consistency between the first medical data and the second medical data is performed. The first medical data includes, as the first input data, information on the treatment site of the patient. The second medical data includes, as the second examination data, treatment image data acquired during the treatment of the patient. Based on the image data taken during treatment, the target area for treatment is identified, and the consistency evaluation is performed by comparing the identified target area with the treatment target area information. method.

11. A method for processing medical data to support patient care, Prepare a first medical data set that includes either the first input data entered by a healthcare professional, the first test data obtained based on data acquired through medical tests, or pre-created standard data. Prepare second medical data, which includes either second input data entered by a healthcare professional, or second test data obtained based on data obtained through medical tests. An evaluation of the consistency between the first medical data and the second medical data is performed. The aforementioned medical examinations include ophthalmological examinations, The examination data obtained by the aforementioned ophthalmic examination includes at least one of the following: intraocular pressure data, optical coherence tomography data, captured image data, strabismus data, heterophoria data, visual acuity data, depth perception data, refractive power data, ocular aberration data, visual field data, corneal morphology data, fundus morphology data, corneal endothelial cell data, axial length data, binocular vision function data, and color vision data. method.

12. A program that causes a computer to process medical data in order to support the treatment of patients, To the aforementioned computer, A step of preparing first medical data which includes either first input data entered by a medical professional, first test data obtained based on data obtained by medical tests, or pre-created standard data. A step of preparing second medical data which includes either second input data entered by a medical professional, or second test data obtained based on data obtained through medical tests, A step of performing an evaluation of the consistency between the first medical data and the second medical data. Make it run, The first medical data includes a first data group consisting of two or more data, and / or the second medical data includes a second data group consisting of two or more data. As a step of performing the evaluation, the computer is made to perform the evaluation of consistency based on the first data group and / or the second data group. The first data set includes standard range information for the examination of a predetermined item and relational information indicating the relationship between the examination of the predetermined item and a predetermined medical procedure. The second data set includes test data obtained by applying the predetermined tests to a patient and medical procedure information indicating the medical procedures performed on the patient. As a step in performing the aforementioned evaluation, the computer will: Based on the relationship information, the determination of whether or not there is a relationship between the examination of the predetermined items and the medical procedure indicated in the medical procedure information is performed. If it is determined that there is no such relationship, the standard range information and the inspection data are compared, and the consistency evaluation is performed based on the results of the comparison. If the aforementioned relationship is determined to exist, the system will perform a comparison between the standard range information and the inspection data, process the results of the comparison based on the relationship information, and perform the aforementioned evaluation of consistency based on the results of the processing. program.

13. A program that causes a computer to process medical data in order to support the treatment of patients, To the aforementioned computer, A step of preparing first medical data which includes either first input data entered by a medical professional, first test data obtained based on data obtained by medical tests, or pre-created standard data. A step of preparing second medical data which includes either second input data entered by a medical professional, or second test data obtained based on data obtained through medical tests, A step of performing an evaluation of the consistency between the first medical data and the second medical data. Make it run, The first medical data includes, as the first input data, information on the treatment site of the patient. The second medical data includes, as the second examination data, treatment image data acquired during the treatment of the patient. As a step in performing the evaluation, the computer is instructed to identify the treatment target area based on the treatment image data, and to perform the evaluation of consistency by comparing the identified target area with the treatment target area information. program.

14. A program that causes a computer to process medical data in order to support the treatment of patients, To the aforementioned computer, A step of preparing first medical data which includes either first input data entered by a medical professional, first test data obtained based on data obtained by medical tests, or pre-created standard data. A step of preparing second medical data which includes either second input data entered by a medical professional, or second test data obtained based on data obtained through medical tests, A step of performing an evaluation of the consistency between the first medical data and the second medical data. Make it run, The aforementioned medical examinations include ophthalmological examinations, The examination data obtained by the aforementioned ophthalmic examination includes at least one of the following: intraocular pressure data, optical coherence tomography data, captured image data, strabismus data, heterophoria data, visual acuity data, depth perception data, refractive power data, ocular aberration data, visual field data, corneal morphology data, fundus morphology data, corneal endothelial cell data, axial length data, binocular vision function data, and color vision data. program.

15. A computer-readable non-temporary recording medium on which a program is recorded that causes a computer to perform processing of medical data to support the treatment of patients, The program is installed on the computer. A step of preparing first medical data which includes either first input data entered by a medical professional, first test data obtained based on data obtained by medical tests, or pre-created standard data. A step of preparing second medical data which includes either second input data entered by a medical professional, or second test data obtained based on data obtained through medical tests, A step of performing an evaluation of the consistency between the first medical data and the second medical data. Make it run, The first medical data includes a first data group consisting of two or more data, and / or the second medical data includes a second data group consisting of two or more data. As a step of performing the evaluation, the program causes the computer to perform the evaluation of consistency based on the first data group and / or the second data group. The first data set includes standard range information for the examination of a predetermined item and relational information indicating the relationship between the examination of the predetermined item and a predetermined medical procedure. The second data set includes test data obtained by applying the predetermined tests to a patient and medical procedure information indicating the medical procedures performed on the patient. As a step in performing the evaluation, the program is provided to the computer, Based on the relationship information, the determination of whether or not there is a relationship between the examination of the predetermined items and the medical procedure indicated in the medical procedure information is performed. If it is determined that there is no such relationship, the standard range information and the inspection data are compared, and the consistency evaluation is performed based on the results of the comparison. If the aforementioned relationship is determined to exist, the system will perform a comparison between the standard range information and the inspection data, process the results of the comparison based on the relationship information, and perform the aforementioned evaluation of consistency based on the results of the processing. Recording medium.

16. A computer-readable non-temporary recording medium on which a program is recorded that causes a computer to perform processing of medical data to support the treatment of patients, The program is installed on the computer. A step of preparing first medical data which includes either first input data entered by a medical professional, first test data obtained based on data obtained by medical tests, or pre-created standard data. A step of preparing second medical data which includes either second input data entered by a medical professional, or second test data obtained based on data obtained through medical tests, A step of performing an evaluation of the consistency between the first medical data and the second medical data. Make it run, The first medical data includes, as the first input data, information on the treatment site of the patient. The second medical data includes, as the second examination data, treatment image data acquired during the treatment of the patient. As a step in performing the evaluation, the program causes the computer to identify the treatment target area based on the treatment image data, and to perform the evaluation of consistency by comparing the identified target area with the treatment target area information. Recording medium.

17. A computer-readable non-temporary recording medium on which a program is recorded that causes a computer to perform processing of medical data to support the treatment of patients, The program is installed on the computer. A step of preparing first medical data which includes either first input data entered by a medical professional, first test data obtained based on data obtained by medical tests, or pre-created standard data. A step of preparing second medical data which includes either second input data entered by a medical professional, or second test data obtained based on data obtained through medical tests, A step of performing an evaluation of the consistency between the first medical data and the second medical data. Make it run, The aforementioned medical examinations include ophthalmological examinations, The examination data obtained by the aforementioned ophthalmic examination includes at least one of the following: intraocular pressure data, optical coherence tomography data, captured image data, strabismus data, heterophoria data, visual acuity data, depth perception data, refractive power data, ocular aberration data, visual field data, corneal morphology data, fundus morphology data, corneal endothelial cell data, axial length data, binocular vision function data, and color vision data. Recording medium.

Images