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

Abstract

To generate data for machine learning to predict the occurrence of an event of a patient with a temporal resolution higher than that of temporal information registered in a medical information system.SOLUTION: A medical information processing method is executed by a computer. The medical information processing method includes: acquiring event information registered in a medical information system and including a patient ID which is identification information on a patient, identification information on an event, and first temporal information indicating the occurrence time of the event; acquiring first medical care information including second temporal information with a temporal resolution higher than that of the first temporal information related to the patient using the patient ID and the first temporal information; and estimating the occurrence time of the event with the temporal resolution higher than that of the first temporal information on the basis of the first medical care information.SELECTED DRAWING: Figure 4

Description

Technical Field

[0001] The present disclosure relates to a medical information processing method, a medical information processing apparatus, and a program.

Background Art

[0002] In recent years, the application of artificial intelligence to the medical field has been promoted. For example, it has been proposed to generate a learning dataset including a plurality of sets of medical records and medical information including patient vital data before the creation of the medical record, perform machine learning using the learning dataset, and generate a learned model (see, for example, Patent Document 1). Thereby, the occurrence of events described in the medical record can be predicted using medical information such as vital data as input. Such an apparatus can be used for predicting the prognosis of a patient and assisting in formulating a treatment plan for a doctor.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] When performing machine learning using retrospective medical data accumulated in a medical information system of a medical institution, the information in a medical record such as an electronic medical record only includes date information of information to be predicted such as the occurrence of an event, and often does not include detailed time information. For example, data such as dialysis transfer and introduction / removal of a respirator, when recorded in an electronic medical record for medical fee billing as receipt data, often includes date information but does not include time information. Also, the record of death often includes information on the date of death but does not include time information.

[0005] On the one hand, the condition of a patient can change suddenly, especially in an ICU (Intensive Care Unit) that has critically ill patients. Therefore, in order to predict and manage changes in the patient's condition, prediction with a higher time resolution than daily is required. However, even if the occurrence of an event is predicted using a trained model generated with daily event information accumulated in the electronic medical record as the target variable, only predictions with a low time resolution can be made, and there is concern that it may be insufficient for use in the medical treatment of patients.

[0006] Therefore, the object of the present disclosure made by focusing on these points is to generate data for machine learning for predicting the occurrence of a patient's event with a higher time resolution than the time information included in the information registered in a medical information system such as an electronic medical record.

Means for Solving the Problems

[0007] A medical information processing method as one aspect of the present disclosure is a medical information processing method executed by a computer, which acquires event information including a patient ID that is identification information of a patient registered in a medical information system, identification information of an event, and first time information indicating the occurrence time of the event, and uses the patient ID and the first time information to acquire first medical information including second time information having a higher time resolution than the first time information related to the patient, and estimates the occurrence time of the event with a higher time resolution than the first time information based on the first medical information.

[0008] As one embodiment, the first medical information includes the patient's vital data.

[0009] As one embodiment, the medical information processing method includes estimating the occurrence time of the event based on the second time information when the value of the data of a predetermined data item of the vital data has a predetermined change.

[0010] As one embodiment, the medical information processing method includes estimating the occurrence time of the event based on the second time information when a data item of the vital data is changed.

[0011] As one embodiment, the first medical information includes drug administration data indicating information on drug administration performed on the patient.

[0012] As one embodiment, the first medical information includes test value data indicating the results of tests performed on the patient.

[0013] As one embodiment, the first medical information includes time-series data generated in a time series based on the second time information, and the medical information processing method includes estimating the occurrence time of the event based on a change in a time interval at which the time-series data is generated.

[0014] As one embodiment, the first medical information includes text data input into an electronic medical record or other medical system and the second time information when the text data is input.

[0015] As one embodiment, the medical information processing method includes estimating the occurrence time of the event based on a combination of two or more types of the first medical information.

[0016] As one embodiment, the event includes a specific event associated with one or more specific data items included in the first medical information, and the medical information processing method includes estimating the occurrence time of the specific event based only on the specific data items.

[0017] As one embodiment, the medical information processing method includes estimating the occurrence time of the event using a first learned model that takes the first medical information as input and outputs the occurrence time of the event.

[0018] As one embodiment, the medical information processing method designates the event as a first event, uses information including the first event and the estimated occurrence time as first event information, and uses the first event information as the first medical information, and includes estimating the occurrence time of a second event different from the first event.

[0019] As one embodiment, the medical information processing method includes obtaining second medical information used as input data of a machine learning model that predicts the occurrence of the event, and associating the second medical information with the estimated occurrence time of the event.

[0020] As one embodiment, the medical information processing method includes associating the second medical information with the occurrence time of the event by adding information indicating the occurrence time of the event to the second medical information.

[0021] As one embodiment, the medical information processing method uses, as teacher data, the event information about a plurality of patients and the second medical information associated with the occurrence time of the event, and generates a second trained model that takes the second medical information as input and outputs the event information using the teacher data.

[0022] As one embodiment, the medical information processing method includes obtaining third medical information of a specific patient, inputting the third medical information into the second trained model, and predicting the occurrence of the event related to the specific patient.

[0023] As one embodiment, the medical information system is an electronic medical record system.

[0024] As one aspect of the present disclosure, a medical information processing apparatus acquires event information including a patient ID which is identification information of a patient registered in a medical information system, identification information of an event, and first time information indicating the occurrence time of the event, and uses the patient ID and the first time information to acquire first medical information including second time information having a higher time resolution than the first time information related to the patient, and includes a control unit that estimates the occurrence time of the event with a higher time resolution than the first time information based on the first medical information.

[0025] As one aspect of the present disclosure, a program causes a computer to execute a process of acquiring event information including a patient ID which is identification information of a patient registered in a medical information system, identification information of an event, and first time information indicating the occurrence time of the event, using the patient ID and the first time information to acquire first medical information including second time information having a higher time resolution than the first time information related to the patient, and estimating the occurrence time of the event with a higher time resolution than the first time information based on the first medical information.

Advantages of the Invention

[0026] The medical information processing method, medical information processing apparatus, and program of the present disclosure estimate the occurrence time of an event based on first medical information including second time information having a higher time resolution than first time information registered in a medical information system. Accordingly, according to the present disclosure, it is possible to generate data for machine learning for predicting the occurrence of a patient event with a higher time resolution than the time information included in the information registered in the medical information system.

Brief Description of the Drawings

[0027]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Embodiments for Carrying Out the Invention

[0028] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings.

[0029] (Configuration of the In-Hospital System) The medical information processing apparatus 11 according to an embodiment of the present disclosure is included in the in-hospital system 10 as shown in FIG. 1. The in-hospital system 10 includes a system that performs machine learning with the event information registered in the medical information system in the hospital as the output (objective variable) and the medical information as the input (explanatory variable) to predict an event that occurs to a patient from the medical information. The medical information system is all devices and systems that handle patient information. The medical information system includes, for example, computers for receipt creation in medical institutions and the like, electronic medical record systems, and systems that support medical affairs and medical treatment such as ordering systems, and computers that hold patient information in some form. Hereinafter, the medical information system will be described as an electronic medical record system.

[0030] In the present application, an event means a major change in the state that has occurred to a patient. An event is information accompanied by a diagnosis by a doctor and / or information on a treatment by a doctor. Events registered in the electronic medical record include the onset of a patient's complication, death, the onset of sepsis or other infectious diseases, shock, the execution of a treatment on the patient, and the like. Complications include cerebral infarction, intracranial hemorrhage, myocardial infarction, diabetes insipidus, hyponatremia, hypernatremia, renal dysfunction, and liver dysfunction. Treatments for patients include the introduction of dialysis and the introduction and removal of a respirator.

[0031] Medical information is data related to a patient's medical treatment. Medical information includes data in an electronic medical record, the patient's vital data, medication administration data, test value data, and other data related to the patient's medical treatment. Vital data includes measurement items such as blood pressure, respiratory rate, heart rate, body temperature, arterial oxygen saturation (SpO2), and urine output. Vital data includes information on the time when the vital data was measured as time information. Medication administration data includes identification information of the medication administered to the patient and information on the dosage. Medication administration data includes information on the time when the medication was administered as time information. Test value data includes data indicating the test results of various test items obtained by various tests including blood tests, electrocardiogram tests, and chest X-ray tests. Test value data includes information on the time when the test was performed as time information. Hereinafter, each individual measurement item of the vital data, each medication of the medication administration data, and each test item of the test value data included in the medical information are referred to as data items. The data items of the medical information can be paraphrased as variables of the medical information.

[0032] The in-hospital system 10 includes, for example, in addition to the medical information processing device 11, a learning device 12, a prediction device 13, an electronic medical record system 14, a vital data management server 15, a medication administration data management server 16, a test value data management server 17, and an information terminal 18. Each device is a computer communicably connected to each other by a network 19. The network 19 can use, for example, a wired LAN (Local Area Network) and a wireless LAN. The computers include PCs (personal computers), PC servers, workstations, general-purpose computers, and other computers.

[0033] The medical information processing device 11 is configured to be able to collect event information from the electronic medical record system 14. The event information includes patient ID which is the identification information of the patient, the identification information of the event, and the date information of the event occurrence date. The date information is the first time information of the present disclosure. Further, the medical information processing device 11 is configured to be able to acquire, together with the time information, the first medical information of the patient on the event occurrence date corresponding to the event information from the vital data management server 15, the drug administration data management server 16, and the test value data management server 17. The time information is the second time information of the present disclosure. The first medical information is the medical information used for estimating the occurrence time of the event. The occurrence time of the event is the information of the event occurrence time and has a higher time resolution than the date information. The medical information processing device 11 may search the time information included in the medical information based on the date information of the event occurrence date and acquire the first medical information including the information of the event occurrence date. The medical information processing device 11 can estimate the occurrence time of the event with a time resolution higher than the daily unit based on the acquired first medical information. The medical information processing device 11 associates the second medical information with the estimated occurrence time of the event. The second medical information is the medical information used for the input data of machine learning. The medical information processing device 11 may be operated by an operator in charge of data collection and processing, etc.

[0034] The first medical information and the second medical information each contain data of one or more data items. The first medical information and the second medical information may be the same information, may be partially overlapping and partially different information, or may be completely different information. As the first medical information, in order to specify the occurrence time of a specific event, data of data items directly related to the event and changing before and after the event are preferentially selected. On the other hand, as the second medical information, data of data items whose causal relationship is presumed with respect to the occurrence of the event are preferentially selected. For example, as the first medical information, data of data items included in vital data containing information on the event date may be selected, and as the second medical information, data of data items included in drug administration data to the patient may be selected. The second medical information may include data on treatment and procedures by a doctor described in the electronic medical record.

[0035] The learning device 12 uses, as teacher data, a large number of event information about a large number of patients and the second medical information associated with the occurrence time of each event, and generates a learned model (second learned model) by performing machine learning using this teacher data. The event information is the target variable of the machine learning, and the second medical information is the explanatory variable of the machine learning. The learning device 12 delivers the generated learned model to the prediction device 13.

[0036] The prediction device 13 predicts an event occurring to a patient using the learned model generated by the learning device 12. In other words, the prediction device 13 predicts the prognosis of the patient. The prediction device 13 acquires the third medical information of a specific patient to be predicted. The third medical information includes medical information of the same data items (variables) as the second medical information. The prediction device 13 inputs the acquired third medical information into the learned model generated by the learning device 12, and outputs event information including the predicted event and its occurrence time.

[0037] The functions of the medical information processing device 11, the learning device 12, and the prediction device 13 may be implemented on the same hardware rather than on separate hardware. For example, the functions of the medical information processing device 11, the learning device 12, and the prediction device 13 may all be realized by a single computer. Also, for example, the functions of the medical information processing device 11 and the learning device 12 may be realized by a single computer that executes a learning phase of machine learning different from the prediction device 13.

[0038] The electronic medical record system 14 manages data of electronic medical records used within a medical institution. The electronic medical record system 14 is included in the medical information system. Information on a patient's event is input into the electronic medical record by a doctor or a nurse. The information on the event input into the electronic medical record system 14 includes the date information of the occurrence of the event, but does not necessarily include time information finer than the date. A part of the patient's medical information may be input into the electronic medical record system 14. Findings of examination, prescriptions of drugs, and details of treatment, etc. input by a doctor in text may be recorded in the electronic medical record system 14 together with time information.

[0039] Other medical systems other than the electronic medical record system may also include text comments input by a doctor or a nurse. For example, other medical systems include systems that manage images such as CT (Computed Tomography) or MRI (Magnetic Resonance Imaging). Comments input into these medical systems may also be registered together with time information entered by the system. The medical information processing device 11 may be configured to be able to acquire information from these other medical systems.

[0040] The vital data management server 15, the medication administration data management server 16, and the test value data management server 17 each manage the vital data, medication administration data, and test value data of each patient. The vital data management server 15, the medication administration data management server 16, and the test value data management server 17 may be configured as databases equipped with a database management system.

[0041] The vital data, medication administration data, and test value data may be managed by a single server, rather than by different server devices as shown in FIG. 1. Also, the vital data, medication administration data, and test value data may be managed by a plurality of different hardware in various combinations. The medical information may be classified and managed in a manner different from the classifications listed here. At least a part of the vital data, medication administration data, and test value data may be stored and managed by the electronic medical record system 14.

[0042] The information terminal 18 is a terminal used by medical staff such as doctors or nurses in a consultation room or the like. The information terminal 18 can use, for example, a PC. The information terminal 18 can display, add, and change the contents of the electronic medical record managed by the electronic medical record system 14. The information terminal 18 can display the data managed by the vital data management server 15, the medication administration data management server 16, and the test value data management server 17. The information terminal 18 can send an instruction for predicting the occurrence of a patient event to the prediction device 13 and obtain the result predicted by the prediction device 13.

[0043] (Configuration of the medical information processing device) As shown in FIG. 2, the medical information processing device 11 includes a control unit 21, a storage unit 22, a communication unit 23, an input unit 24, and an output unit 25. Note that the learning device 12, the prediction device 13, the electronic medical record system 14, the vital data management server 15, the medication administration data management server 16, and the test value data management server 17 may be configured to include a control unit, a storage unit, a communication unit, an input unit, and an output unit, similar to the medical information processing device 11.

[0044] The control unit 21 includes at least one processor. The processor includes a general-purpose processor such as a CPU (central processing unit), or a dedicated processor specialized for specific processing. The control unit 21 may include an application specific integrated circuit (ASIC), a digital signal processor (DSP), a programmable logic device (PLD), a field-programmable gate array (FPGA), or any combination thereof. The control unit 21 may include a memory built into the processor or a memory independent of the processor. The control unit 21 executes processes related to the operation of the medical information processing device 11 while controlling each part of the medical information processing device 11.

[0045] The storage unit 22 includes at least one semiconductor memory, at least one magnetic memory, at least one optical memory, or any combination thereof. The semiconductor memory is, for example, a RAM (random access memory) or a ROM (read only memory). The RAM is, for example, an SRAM (static random access memory) or a DRAM (dynamic random access memory). The ROM is, for example, an EEPROM (electrically erasable programmable read only memory). The storage unit 22 functions as, for example, a main storage device, an auxiliary storage device, or a cache memory. Stored in the storage unit 22 are data used for the operation of the medical information processing device 11 and data obtained by the operation of the medical information processing device 11.

[0046] The communication unit 23 includes at least one communication interface. The communication interface is, for example, an interface corresponding to a mobile communication standard such as a LAN interface, LTE (Long Term Evolution), 4G (4th generation) standard, or 5G (5th generation) standard, or an interface corresponding to a short-range wireless communication standard such as Bluetooth (registered trademark). The communication unit 23 receives data used for the operation of the medical information processing device 11 and transmits data obtained by the operation of the medical information processing device 11. The medical information processing device 11 can acquire event information included in an electronic medical record (medical information system) and first medical information including vital data, drug administration data, and test value data via the communication unit 23.

[0047] The input unit 24 includes at least one input interface. The input interface is, for example, a physical key, a capacitive key, a pointing device, a touch screen provided integrally with a display, an imaging device such as a camera, or a microphone. The input unit 24 receives an operation for inputting data used for the operation of the medical information processing device 11. The input unit 24 receives an operation by an operator on the event information and the first medical information. Instead of being provided in the medical information processing device 11, the input unit 24 may be connected to the medical information processing device 11 as an external input device.

[0048] The output unit 25 includes at least one output interface. The output interface is, for example, a display or a speaker. The display is, for example, an LCD (liquid crystal display) or an organic EL (electro luminescence) display. The output unit 25 outputs data obtained by the operation of the medical information processing device 11. Instead of being provided in the medical information processing device 11, the output unit 25 may be connected to the medical information processing device 11 as an external output device.

[0049] The functions of the medical information processing apparatus 11 are realized by executing the program according to this embodiment on a processor as the control unit 21. That is, the functions of the medical information processing apparatus 11 are realized by software. The program causes a computer to execute the operations of the medical information processing apparatus 11, thereby making the computer function as the medical information processing apparatus 11. That is, the computer functions as the medical information processing apparatus 11 by executing the operations of the medical information processing apparatus 11 according to the program.

[0050] The program can be stored in a non-transitory computer-readable medium. The non-transitory computer-readable medium is, for example, a flash memory, a magnetic recording device, an optical disk, a magneto-optical recording medium, or a ROM. The distribution of the program can be performed, for example, by selling, transferring, or lending a portable medium such as an SD (Secure Digital) card storing the program, a DVD (digital versatile disc), or a CD-ROM (compact disc read only memory). The program may be stored in the server storage and distributed by transferring the program from the server to other computers. The program may be provided as a program product.

[0051] The computer stores, once, a program stored in a portable medium or a program transferred from a server in a main storage device, for example. Then, the computer reads the program stored in the main storage device with a processor and executes processing according to the read program with the processor. The computer may directly read a program from a portable medium and execute processing according to the program. The computer may sequentially execute processing according to the received program each time a program is transferred from a server to the computer. Processing may be executed by a so-called ASP (application service provider) type service that realizes functions only by execution instructions and result acquisition without transferring a program from a server to a computer. A program includes information for use in processing by an electronic computer and things conforming to the program. For example, data that is not a direct instruction to a computer but has a property of defining processing of the computer corresponds to "things conforming to the program."

[0052] Some or all functions of the medical information processing apparatus 11 may be realized by a programmable circuit or a dedicated circuit as the control unit 21. That is, some or all functions of the medical information processing apparatus 11 may be realized by hardware.

[0053] (Overall process of machine learning) With reference to FIG. 3, the process of machine learning executed in the in-hospital system 10 will be described. The process of machine learning includes a flow during learning and a flow during prediction.

[0054] The flow during learning includes steps of "data collection", "preprocessing", "association", "feature extraction", "learning", "model output", and "evaluation".

[0055] "Data collection" includes the process of collecting event information from the electronic medical record system 14, and the process of collecting medical information from the vital data management server 15, the medication administration data management server 16, and the test value data management server 17. The medical information is used for the first medical information and / or the second medical information. The operator can collect this information using the medical information processing device 11.

[0056] "Preprocessing" is the process of organizing the collected event information and medical information to facilitate machine learning. "Preprocessing" includes data cleansing, data integration, and data transformation. Since these preprocessing steps are well-known in the field of machine learning, the description is omitted. The medical information processing device 11 may include software that assists the operator in performing preprocessing.

[0057] "Association" is the process of estimating the occurrence time of an event with a higher time resolution than daily based on the medical information, and associating the medical information with the estimated occurrence time of the event. "Association" is a characteristic process of the medical information processing method performed by the medical information processing device 11 of the present disclosure. The estimation of the occurrence time of the event is performed based on the first medical information extracted from the overall medical information. The association between the medical information and the occurrence time of the event can be performed, for example, by providing a flag area indicating the occurrence of the event in the medical information collected from each device and setting the flag area of the event occurrence time to a predetermined value. Also, "association" may be performed by adding time information to the event information. "Association" can be paraphrased as labeling.

[0058] "Feature extraction" is a process of extracting the second medical information necessary for machine learning from the data of medical information and determining the second medical information to be input to the machine learning. The extraction of the second medical information is performed by extracting the data items of medical information that are estimated to affect the occurrence of an event. "Feature extraction" may be executed by an operator using the learning device 12. Feature extraction may be automatically executed by the learning device 12 based on the associated event information and medical information without depending on the operator.

[0059] "Learning" is a process of performing machine learning using teacher data with event information as the target variable and the second medical information extracted from the entire medical information as the explanatory variable, and generating a learned model. For machine learning, methods such as a recurrent neural network (RNN), random forest, gradient boosting, or support vector machine (SVM) can be used. "Learning" is executed by the learning device 12. The learning device 12 may execute "feature extraction" and "learning" simultaneously.

[0060] "Model output" is a process of outputting a learned model for the prediction flow. The learning device 12 may transmit the learned model to the prediction device 13 via the network 19.

[0061] "Evaluation" is a process of evaluating a learned model. The learning device 12 evaluates the prediction accuracy of the learned model using a part of the data extracted by feature extraction as test data. If the prediction accuracy is low, the learning device 12 may return to "feature extraction" to change the data items used for machine learning. If the prediction accuracy is low, the learning device 12 may also return to the processing of "preprocessing" or "association".

[0062] The prediction flow includes the steps of "input", "feature extraction", "prediction", and "result output". The prediction process may be started when a doctor, nurse, or professional operator sends an instruction from a terminal device to the prediction device 13.

[0063] "Input" is the process of acquiring the medical information of the patient to be predicted. When the prediction device 13 receives an instruction to predict the occurrence of an event related to a patient from the information terminal 18, it acquires the medical information of that patient from the electronic medical record system 14, the vital data management server 15, the drug administration data management server 16, the test value data management server 17, etc.

[0064] "Feature extraction" is the process of extracting data of the third medical information to be used for prediction from the medical information of the patient. The prediction device 13 extracts data of the same data items as the second medical information extracted in "feature extraction" in the learning phase from the input medical information as input data to the learned model. Alternatively, as part of the third medical information, information on procedures or medications that have not yet been executed may be input from the information terminal 18 in order to predict the effects of procedures or medications, etc.

[0065] "Prediction" is the process of inputting the third medical information as input data into the learned model generated in the learning phase to predict the occurrence of an event. The prediction device 13 executes "prediction".

[0066] In "result output", the prediction device 13 outputs the prediction result. The prediction device 13 outputs the predicted event and the predicted occurrence time of the event. The prediction device 13 may output the prediction result to the display device of the information terminal 18 operated by the doctor. Based on the prediction result, the doctor can consider the medical treatment policy.

[0067] (Processing flow of the medical information processing device) Next, with reference to FIGS. 4 and 5, among the processes executed by the medical information processing apparatus 11, the processing content related to "association", which is a feature of the present disclosure, will be described. The processes of the flowcharts in FIGS. 4 and 5 may be executed while the operator operates the medical information processing apparatus 11 using the input unit 24 and the output unit 25 of the medical information processing apparatus 11. Further, the processes of the flowcharts in FIGS. 4 and 5 may be partially automated or fully automated.

[0068] First, the control unit 21 of the medical information processing apparatus 11 acquires event information including a patient ID, which is identification information of a patient registered in the electronic medical record system 14 of the patient (medical information system), identification information of an event, and date information of the occurrence date of the event from the electronic medical record system 14 (step S101). The control unit 21 may store the acquired event information in the storage unit 22. This process may be executed in the "data collection" step of FIG. 3.

[0069] Next, the control unit 21 acquires first medical information related to the above event using the patient ID and the date information (step S102). The control unit 21 can search for medical information using the patient ID and the date information as search conditions. The control unit 21 may acquire medical information in which the date information included in the time information matches the date information included in the patient's event information from the entire medical information. For example, the first medical information to be acquired includes time-series vital data, drug administration data, or test value data of the patient on the day when the event occurred to the patient. The control unit 21 may store the acquired first medical information in the storage unit 22. The medical information processing apparatus 11 may search the medical information acquired in the "data collection" step of FIG. 3 to acquire the first medical information. Alternatively, the medical information processing apparatus 11 may directly search the vital data management server 15, the drug administration data management server 16, and the test value data management server 17 to acquire the first medical information.

[0070] Next, for each set of respective event information and first medical information, the control unit 21 estimates the occurrence time of the event based on the first medical information (step S103). That is, although the event information often contains only the information of the date when the event occurred, the control unit 21 estimates the time when the event occurred with a higher time resolution than the date by using the first medical information.

[0071] A method for estimating the occurrence time of an event in the medical information processing apparatus 11 will be described with reference to the flowchart of FIG. 5.

[0072] First, the control unit 21 extracts first medical information that is related to the event or is estimated to have a high relevance to the event based on the content of the event (step S201). For example, the control unit 21 acquires time-series vital data of the date when the event occurred. Vital data is considered to be the most prioritized information that can be used for estimating the occurrence time of the event.

[0073] Next, the control unit 21 extracts a specific change from the extracted first medical information (step S202). The specific change includes a change in the value of data of a specific data item corresponding to the event, a change in the data acquisition frequency, and a change in the data item.

[0074] The change in the value of data of a specific data item includes cases where the value of data of a specific data item included in the medical information becomes an abnormal value, rises rapidly, and drops rapidly, etc. For example, when the medical information is vital data, the change in the value of data of a specific data item includes cases where the patient's blood pressure, heart rate, or body temperature exceeds the normal value range and falls below the normal value range. When the first medical information is drug administration data, the change in the value of data of a specific data item includes cases where the administration amount of a specific drug has increased.

[0075] Changes in the data acquisition frequency include changes in the time interval at which data is acquired. For example, changes in the data acquisition frequency include changes in the measurement frequency of vital data. For example, for a patient with stable condition, vital data may be measured every two to three hours. In contrast, when the patient's condition deteriorates, vital data may be measured at shorter intervals. Therefore, it can be estimated that an event has occurred because the time interval at which vital data is measured has become shorter.

[0076] Changes in data items include the fact that data for data items that were not acquired before are now being acquired. For example, end-tidal carbon dioxide partial pressure (EtCO2) is an index for evaluating whether a patient can ventilate. An EtCO2 monitor that measures EtCO2 is used to manage the respiratory state of a patient under ventilator management. Therefore, the timing when EtCO2 is acquired as a data item of vital data is estimated to be the timing when the ventilator is introduced.

[0077] When the control unit 21 extracts a specific change from the first medical information in step S202, it acquires information on the time when the specific change occurred (step S203).

[0078] The control unit 21 estimates the time when the event occurred from the information on the occurrence time of the acquired specific change (step S204). In some events, the occurrence time of the specific change acquired in step S203 is estimated to be equal to the time when the event occurred. Also, in some other events, considering that there is a time lag between the occurrence time of the specific change acquired in step S203 and the time when the event occurred, the occurrence time of the event is estimated.

[0079] When the occurrence time of an event can be estimated in step S204 (step S205: Yes), the control unit 21 returns to the flowchart of FIG. 4 and continues the process. When the occurrence time of the event cannot be estimated in step S204 (step S205: No), the control unit 21 extracts other first medical information that is highly likely to be related to the event (step S206) and repeats the processes after step S202.

[0080] For example, as shown in FIG. 6, the control unit 21 may first acquire vital data. When the occurrence time of the event cannot be estimated only from the vital data, the control unit 21 sequentially acquires drug administration data, test value data, and other data as other first medical information and may execute the processes from step S202 to S205. However, the order of acquiring the first medical information is not limited to this order. The storage unit 22 stores the order of acquiring the first medical information optimized according to the event, and the control unit 21 may acquire the first medical information based on this order.

[0081] The other data in FIG. 6 may include, for example, text data indicating a doctor's findings, prescriptions, and treatments included in an electronic medical record or other medical information systems, and information on the input time thereof. The control unit 21 may use text mining technology to acquire information highly relevant to the event from the text information included in the electronic medical record and estimate the occurrence time of the event from the input time of the text.

[0082] The control unit 21 may estimate the occurrence time of the event based on a combination of at least two or more types of data included in the vital data, drug administration data, test value data, and other data.

[0083] After estimating the occurrence time of an event in step S205 of FIG. 5, the control unit 21 associates the second medical information with the occurrence time of the event as shown in FIG. 4 (step S104). For example, the control unit 21 can associate the second medical information with the occurrence time of the event by attaching a flag indicating the occurrence of the event to the data of the occurrence time of the event among the time-series second medical information. Alternatively, the control unit 21 can attach time information to the event information itself and indirectly associate it with the second medical information. Data among the plurality of second medical information used as input (explanatory variables) in the subsequent learning process may be associated with the event information (objective variable).

[0084] (Specific Examples of Estimation and Association of Event Occurrence Time) With reference to FIG. 7, an example of a method for estimating the event occurrence time and its association with the second medical information will be described. The control unit 21 of the medical information processing apparatus 11 acquires the identification information of the event that occurred to the patient, the patient ID which is the identification information of the patient, and the date information of the event occurrence date from the information of the electronic medical record (medical information system). As the identification information of the event, the name of the event can be used. In the case of the example in FIG. 7, the control unit 21 acquires event information indicating that mechanical ventilation was introduced to the patient with patient ID 123 on March 1, 2021.

[0085] The control unit 21 can refer to various criteria for estimating the time when the event occurred. Such criteria may be stored in the storage unit 22 in advance. For example, in the case of the introduction of a ventilator, the respiratory rate is medical information related to the introduction of the ventilator. There is a known ventilator introduction criterion that a ventilator should be introduced when the respiratory rate is 5 times per minute or less or 35 times per minute or more. Therefore, when the respiratory rate per minute becomes 35 times or more, it is estimated that the doctor will introduce a ventilator. As the criteria, criteria defined by countries, public institutions, etc., criteria defined within a medical institution, or criteria derived from past medical information in a medical institution can be used.

[0086] The control unit 21 searches for vital data as the first medical information using the patient ID and the date information of the electronic medical record, and acquires vital data including the data on the event occurrence date of the patient. The control unit 21 refers to the value of the respiratory rate included in the vital data, and extracts data in which the respiratory rate has changed from less than 35 times per minute to 35 times or more per minute. In the illustrated example, the patient's respiratory rate is 35 times at 12:30.

[0087] The control unit 21 estimates the introduction time of the ventilator in consideration of the time when the respiratory rate of the vital data exceeds the standard, the time required to introduce the ventilator, and other medical information. In the illustrated example, the introduction time of the ventilator is estimated to be 12:38.

[0088] Next, the control unit 21 associates the second medical information with the occurrence time of the event. Here, assuming that the data on the respiratory rate of the vital data is also included in the second medical information, the data indicating the respiratory rate is shown. As shown in FIG. 8, the control unit 21 may add a flag area for associating the vital data on the respiratory rate stored in the storage unit 22 with the data in the vital data and the occurrence time of the event. For example, the control unit 21 assigns a flag indicating that the ventilator has been introduced to the vital data measured at each time. In FIG. 8, the data in the row where the column of "ventilator introduction flag" is "1" indicates the data at the timing when it is estimated that the ventilator has been introduced. If the vital data at the timing when it is estimated that the ventilator has been introduced has not been measured, the control unit 21 may add it by interpolating the data at the time when it is estimated that the ventilator has been introduced. Similarly to the above, the data included in the second medical information other than the respiratory rate may be associated with the occurrence time of the event.

[0089] (Examples of events and medical information related to the events) FIG. 9 is a diagram showing an event and data items of first medical information related to the event. The first medical information related to several events will be described below. The control unit 21 may estimate the time when an event occurred using only one or a plurality of specific data items of the first medical information corresponding to the event.

[0090] For cerebral infarction and intracranial hemorrhage, it is common to perform imaging such as a head CT for a definite diagnosis. For example, in a head CT examination, the blood flow in the brain is observed by imaging the brain while injecting a contrast agent. Therefore, for the events of the onset of cerebral infarction and the onset of intracranial hemorrhage, the time of occurrence of the event can be estimated by using the data of the time when the contrast agent was administered included in the drug administration data.

[0091] Also, for example, in the diagnosis of myocardial infarction, an electrocardiogram examination, a blood test, and an imaging examination are performed. In the case of myocardial infarction, a waveform peculiar to myocardial infarction appears in the electrocardiogram. Also, when myocardial infarction occurs, an increase in various enzymes such as troponin T and CK-MB in the blood is observed. Furthermore, in a chest X-ray examination, images such as pulmonary congestion or cardiomegaly are taken. Therefore, when these test value data are included as data items of the first medical information, it is estimated that myocardial infarction occurred before the test was performed.

[0092] Also, when the event is death, the time of death can be estimated from vital data such as blood pressure, SpO2, respiratory rate, and heart rate.

[0093] As an example, the control unit 21 can estimate the time when the patient died from only the blood pressure data. FIG. 10 shows an example of time-series data of systolic and diastolic blood pressure. The time when the blood pressure value dropped near 0 can be estimated as the time when the patient died. Similarly, the control unit 21 can estimate the time of death independently from the respiratory rate and the heart rate.

[0094] If, unlike the above, the time of death cannot be determined solely from the decrease in blood pressure value, the control unit 21 may estimate the time when the event of death occurred in combination with other vital data such as respiratory rate or heart rate.

[0095] When the event is sepsis, it can be determined that there is a high possibility of the onset of sepsis when two of the three requirements of having impaired consciousness, having a systolic blood pressure of 100 mmHg or less, and having a respiratory rate of 22 or more times per minute are met. Furthermore, for a definitive diagnosis of sepsis, imaging diagnoses such as chest X-ray examination and CT examination are performed. Therefore, blood pressure and respiratory rate included in vital data, chest X-ray examination results included in examination value data, and contrast agents for CT examination included in drug administration data can be used as data items of the first medical information related to sepsis.

[0096] (Estimation of the occurrence time of an event using machine learning) For a specific event, if there is a set of information on the occurrence time of a certain number or more of accurate or highly accurate events and medical information, it is also possible to use machine learning in the process of estimating the occurrence time of the event. The accurate event occurrence time may be input into an electronic medical record or other medical information system. Also, the time of introduction of a respirator can be predicted with a certain degree of accuracy from the timing when measurement items of vital information related to respiration are added. High accuracy regarding the occurrence time of an event means that the recorded occurrence time has a higher time resolution than the daily unit, and more preferably, it means that the time resolution is higher than the hourly unit.

[0097] The medical information processing device 11 or another dedicated device may use medical information highly relevant to the occurrence of an event as input information, execute machine learning with the occurrence time of the event as the output, and generate a learned model for estimating the event occurrence time (the first learned model). The generated learned model can be stored in the storage unit 22. For an event with an unknown exact occurrence time, the control unit 21 can input data of first medical information including the same data items as the medical information during learning into the learned model for estimating the event occurrence time, thereby enabling the estimation of the time when the event occurred.

[0098] As described above, the medical information processing device 11 may utilize machine learning in the same manner as the prediction device 13. However, the prediction performed by the prediction device 13 in FIG. 1 predicts the occurrence of future events from the patient's medical information, whereas the time estimation performed by the medical information processing device 11 in the association process estimates the time when a specific event that has already occurred occurred, which is different in this regard.

[0099] (Estimation of the occurrence time of another event based on the estimated occurrence time of the event) The control unit 21 can use the event information (first event information) of the event (first event) whose occurrence time has been estimated as first medical information for estimating the occurrence time of another event (second event). For example, the control unit 21 can set the introduction of a ventilator as the first event and death as the second event. That is, the control unit 21 can estimate the introduction time of the ventilator, use the information including the estimated introduction time of the ventilator as the first medical information, and further estimate the time when the patient dies.

[0100] As described above, according to this embodiment, based on the first medical information, the occurrence time of an event is estimated with a higher time resolution than the daily unit, and the second medical information is associated with the estimated occurrence time of the event. Thereby, it is possible to generate data for machine learning that associates the occurrence of a patient event with the second medical information with a higher time resolution than the daily unit. By using this data for machine learning, the learning device 12 can perform machine learning to generate a learned model. Further, when the prediction device 13 inputs the third medical information, which is the medical information of the patient for whom prediction is to be made, to the generated learned model, it becomes possible to predict the occurrence of a patient event with a higher time resolution than the daily unit.

[0101] (Example of dispersedly arranging each device) In the above embodiment, the medical information processing device 11, the learning device 12, the prediction device 13, the electronic medical record system 14, the vital data management server 15, the drug administration data management server 16, and the test value data server are assumed to be located within a medical institution. However, these devices can be dispersedly arranged at geographically distant locations.

[0102] In the example shown in FIG. 11, the medical information processing device 11 and the learning device 12 may be located within the base of the service provider 30 that provides the learned model. In order to improve the accuracy of machine learning, it is preferable that there is a large amount of learning data including medical information and event information. For this reason, it may be preferable for the service provider 30 to collect information from a plurality of medical institutions. The service provider 30 may be an operator that provides the service or one of a plurality of medical institutions that adopt this system. The system of the service provider 30 and the in - system 10 of a plurality of medical institutions may be connected by a wide - area communication means such as a dedicated line, the Internet, or a VPN (Virtual Private Network).

[0103] Further, the test value data may be managed by a test value data management server 41 located in a data center 40 of the operator that provides the test device, rather than by a server within the in-hospital system 10. In this case, the functions of the test value data management server 41 may be provided as a cloud service. The operator's test value data management server 41 can manage the test value data of multiple medical institutions. A medical institution can use the test value data management server 41 as if it were located within the medical institution. The data center 40 and the in-hospital systems 10 of multiple medical institutions may be connected by a wide-area communication means such as a dedicated line, the Internet, or a VPN (Virtual Private Network).

[0104] The arrangement of the devices shown in FIG. 11 is an example, and each device can be arranged in various ways. Also, the functions of each device and the medical information managed by each device can be divided or integrated in various ways. A medical institution can manage other medical data using an external cloud service instead of or in addition to the test value data.

[0105] Embodiments according to the present disclosure have been described based on the drawings and examples. However, it should be noted that those skilled in the art can easily make various modifications or changes based on the present disclosure. Therefore, it should be noted that these modifications or corrections are included in the scope of the present disclosure. For example, functions included in each component or each step can be rearranged so as not to be logically contradictory, and a plurality of components or steps can be combined into one or divided. Although the embodiments according to the present disclosure have been described mainly with respect to apparatuses, the embodiments according to the present disclosure can also be realized as a method including steps executed by each component of the apparatus. The embodiments according to the present disclosure can also be realized as a method, program, or storage medium recording the program executed by a processor included in the apparatus. It should be understood that these are also included in the scope of the present disclosure. Also, the specific events and medical information described in the above embodiments are merely examples. The present disclosure can be applied to various events and medical information.

[0106] In the above embodiment, the date information of the occurrence of an event is used as the first time information, and the time information having a higher time resolution than the daily unit included in the medical information is used as the second time information. However, the first time information and the second time information of the present disclosure are not limited to this combination. For example, when the occurrence time of an event registered in a medical information system is specified in time units and the medical information related to the event is specified in minute units, the information on the occurrence time of the event in time units can be used as the first time information, and the time information of the medical information in minute units can be used as the second time information.

Description of Reference Numerals

[0107] 10 In-hospital system 11 Medical information processing device 12 Learning device 13 Prediction device 14 Electronic medical record system (medical information system) 15 Vital data management server 16 Medication administration data management server 17 Inspection value data management server 18 Information terminal 19 Network 21 Control unit 22 Memory unit 23 Communication unit 24 Input unit 25 Output unit 30 Service provider 40 Data center 41 Inspection value data management server

Claims

1. A medical information processing method executed by a computer, comprising: obtaining event information including a patient ID, which is identification information of a patient registered in a medical information system, identification information of an event, and first time information indicating the occurrence time of the event; using the patient ID and the first time information to retrieve first medical information that includes second time information having a higher time resolution than the first time information of the patient and is highly relevant to the event from medical information including time information, by searching for medical information related to the patient in which the date included in the time information matches the date at the occurrence time of the event; estimating the occurrence time of the event with a higher time resolution than the first time information based on the second time information included in the first medical information; A medical information processing method.

2. The medical information processing method according to claim 1, wherein the first medical information includes vital data of the patient.

3. Extracting a specific change in the data value of a predetermined data item of the vital data from the first medical information that is highly relevant to the event, and obtaining the occurrence time at which the specific change occurred as the second time information, and based on the obtained occurrence time, estimating the occurrence time of the event by considering that the occurrence time is equal to the occurrence time of the event or the time lag from the occurrence time to the occurrence time of the event. The medical information processing method according to claim 2.

4. The event is a deterioration in the patient's condition, The medical information processing method according to claim 2, wherein the occurrence time of the event is estimated based on the second time information when the measured time interval of the vital data becomes shorter.

5. The medical information processing method according to any one of claims 1 to 4, wherein the first medical information includes medication administration data indicating information on medication administration performed on the patient.

6. The medical information processing method according to any one of claims 1 to 5, wherein the first medical information includes test value data indicating the results of tests performed on the patient.

7. The medical information processing method according to any one of claims 1 to 6, wherein the first medical information includes text data input into an electronic medical record or other medical system and the second time information when the text data was input.

8. The first medical information includes specific data items that are directly related to the event and change before and after the event, obtains the occurrence time when the change in the specific data item occurs as the second time information, and estimates the occurrence time of the specific event based on the obtained occurrence time. The medical information processing method according to any one of claims 1 to 7.

9. Regarding the event as the first event, regarding the information including the first event and the estimated occurrence time as the first event information, using the first event information as the first medical information, and estimating the occurrence time of a second event different from the first event. The medical information processing method according to any one of claims 1 to 8.

10. Obtains second medical information used as input data for a machine learning model that predicts the occurrence of the event, and associates the second medical information with the estimated occurrence time of the event. The medical information processing method according to any one of claims 1 to 9.

11. By adding information indicating the occurrence time of the event to the second medical information, the second medical information is associated with the occurrence time of the event. The medical information processing method according to claim 10.

12. Regarding the event information about a plurality of patients and the second medical information associated with the occurrence time of the event as teacher data, and using the teacher data to generate a second trained model that takes the second medical information as input and outputs the event information. The medical information processing method according to claim 10 or 11.

13. In the generation of the second trained model, the second medical information to be input is extracted by extracting data items of medical information that are estimated to affect the occurrence of the event. Receives an instruction to predict the occurrence of the event regarding a specific patient, and obtains the medical information of the specific patient. As the third medical information of the specific patient, obtains data of the same data items as the second medical information extracted in the generation of the second trained model from the obtained medical information, inputs the third medical information into the second trained model, and predicts the occurrence of the event regarding the specific patient. The medical information processing method according to claim 12.

14. The medical information system is an electronic medical record system, and the medical information processing method according to any one of claims 1 to 13.

15. An event information including a patient ID which is identification information of a patient registered in a medical information system, identification information of an event, and first time information indicating the occurrence time of the event is acquired, and by using the patient ID and the first time information, searching for medical information related to the patient in medical information including time information, wherein the date included in the time information matches the date at the occurrence time of the event, to obtain first medical information including second time information having a higher time resolution than the first time information related to the patient and having a high relevance to the event, and a control unit that executes a process of estimating the occurrence time of the event with a higher time resolution than the first time information based on the second time information included in the first medical information A medical information processing apparatus including the same.

16. An event information including a patient ID which is identification information of a patient registered in a medical information system, identification information of an event, and first time information indicating the occurrence time of the event is acquired, and by using the patient ID and the first time information, searching for medical information related to the patient in medical information including time information, wherein the date included in the time information matches the date at the occurrence time of the event, to obtain first medical information including second time information having a higher time resolution than the first time information related to the patient and having a high relevance to the event, and a program for causing a computer to execute a process of estimating the occurrence time of the event with a higher time resolution than the first time information based on the second time information included in the first medical information

Images