WO2022019514A1 - Apparatus, method, and computer-readable recording medium for decision-making in hospital - Google Patents

Apparatus, method, and computer-readable recording medium for decision-making in hospital Download PDF

Info

Publication number
WO2022019514A1
WO2022019514A1 PCT/KR2021/008298 KR2021008298W WO2022019514A1 WO 2022019514 A1 WO2022019514 A1 WO 2022019514A1 KR 2021008298 W KR2021008298 W KR 2021008298W WO 2022019514 A1 WO2022019514 A1 WO 2022019514A1
Authority
WO
WIPO (PCT)
Prior art keywords
patient
hospital
information
medical
decision
Prior art date
Application number
PCT/KR2021/008298
Other languages
French (fr)
Korean (ko)
Inventor
김영학
허신영
전태준
안임진
조용현
채정우
강희준
허재희
Original Assignee
주식회사 라인웍스
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 주식회사 라인웍스 filed Critical 주식회사 라인웍스
Publication of WO2022019514A1 publication Critical patent/WO2022019514A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H40/00—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
    • G16H40/20—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the management or administration of healthcare resources or facilities, e.g. managing hospital staff or surgery rooms
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0633—Workflow analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0637—Strategic management or analysis, e.g. setting a goal or target of an organisation; Planning actions based on goals; Analysis or evaluation of effectiveness of goals
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/10—Office automation; Time management
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/10—Office automation; Time management
    • G06Q10/109—Time management, e.g. calendars, reminders, meetings or time accounting
    • G06Q10/1093—Calendar-based scheduling for persons or groups
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H10/00—ICT specially adapted for the handling or processing of patient-related medical or healthcare data
    • G16H10/60—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for patient-specific data, e.g. for electronic patient records
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/50—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for simulation or modelling of medical disorders

Definitions

  • the present invention relates to a hospital decision-making apparatus, method, computer-readable recording medium and computer program.
  • An object of the present invention is to provide a hospital decision-making apparatus, a method, a computer-readable recording medium, and a computer program.
  • this hospital decision-making device predicts a patient's medical event through a deep learning learning model, and comprehensively considers the predicted medical event, the patient's medical information, and the hospital's state information to make the hospital's decision. and the like may be included in the problems to be solved by the present invention.
  • a hospital decision-making apparatus includes an input/output unit for receiving medical treatment information about a patient; a memory in which medical information for a plurality of patients is converted into data of a predetermined structure format and stored; and a processor electrically connected to the memory, wherein the processor receives the data in the predetermined structural format as an input, and learns by answering at least one medical event that may occur for the plurality of patients as an answer.
  • a doctor of the hospital predicts at least one medical event for the patient through a deep learning learning model, and considers the predicted at least one medical event for the patient, medical treatment information for the patient, and state information of the hospital decision can be made.
  • the medical treatment information for the patient includes medical treatment information according to the time when the patient first visits the hospital, medical information according to the time when the patient visits the hospital after the initial visit to the hospital, and other information other than the hospital It may include at least one of the medical information obtained from the hospital.
  • the medical event may include a mortality rate within a predetermined period after the patient's discharge, a readmission rate within a predetermined period after the patient's discharge, a period from the time the patient reads into the hospital to a discharge time, and within a predetermined period after the patient's discharge at least one of the incidence of heart disease.
  • the decision-making of the hospital may include determining the patient's treatment schedule, determining the bed assignment of the patient, and determining the operating room schedule of the patient.
  • the state information of the hospital includes the number of treatment rooms in the hospital, the number of treatment rooms used by each medical staff of the hospital, outpatient support personnel information for each medical staff, session opening status information for each day of the other medical staff in the same department of the hospital, and the hospital's It may include at least one of bed information in a hospital room, patient status information of the hospital, status information for each department, the type and number of operating rooms in the hospital, schedule information for each operating room, and estimated time required for each operation of the patient. .
  • the decision-making of the hospital includes determining the patient's medical schedule, and determining the patient's medical care ranking based on the medical treatment information for the patient and the predicted patient's medical event,
  • the patient's treatment schedule may be determined by further considering the determined patient's treatment order and medical staff information of the hospital.
  • the medical staff information of the hospital includes the medical treatment schedule information for each medical staff, the information by the patient type, the time information generated by the patient's treatment, the schedule information of at least one of the surgery, the procedure, and the research by the medical staff, and the medical staff It may include at least one of performance information and information on whether or not a medical staff has been placed in the hospital.
  • the decision-making of the hospital includes determining the bed assignment of the patient, and the processor determines the bed assignment rank of the patient based on the medical event information of the patient and the predicted patient, The bed assignment of the patient may be determined by further considering the determined bed assignment order of the patient.
  • a hospital decision-making method includes converting medical treatment information for a plurality of patients into data of a predetermined structural format, and generating data in the predetermined structural format as an input for the plurality of patients Learning a deep learning learning model with at least one possible medical event as the correct answer, receiving medical treatment information about a patient, and predicting at least one medical event for the patient through the deep learning learning model and performing the decision-making of the hospital in consideration of the predicted at least one medical event for the patient, medical treatment information for the patient, and state information of the hospital.
  • the medical treatment information for the patient includes medical treatment information according to the time when the patient first visits the hospital, medical information according to the time when the patient visits the hospital after the initial visit to the hospital, and other information other than the hospital It may include at least one of the medical information obtained from the hospital.
  • the medical event may include a mortality rate within a predetermined period after the patient's discharge, a readmission rate within a predetermined period after the patient's discharge, a period from the time the patient reads into the hospital to a discharge time, and within a predetermined period after the patient's discharge at least one of the incidence of heart disease.
  • the decision-making of the hospital may include determining the patient's treatment schedule, determining the bed assignment of the patient, and determining the operating room schedule of the patient.
  • the state information of the hospital includes the number of treatment rooms in the hospital, the number of treatment rooms used by each medical staff of the hospital, outpatient support personnel information for each medical staff, session opening status information for each day of the other medical staff in the same department of the hospital, and the hospital's It may include at least one of bed information in a hospital room, patient status information of the hospital, status information for each department, the type and number of operating rooms in the hospital, schedule information for each operating room, and estimated time required for each operation of the patient. .
  • the decision-making of the hospital includes determining the treatment schedule of the patient, and the performing of the decision-making of the hospital includes the treatment of the patient based on the treatment information about the patient and the predicted medical event of the patient.
  • the treatment order may be determined, and the treatment schedule of the patient may be determined by further considering the determined treatment order of the patient and the medical staff information of the hospital.
  • the medical staff information of the hospital includes the medical treatment schedule information for each medical staff, the information by the patient type, the time information generated by the patient's treatment, the schedule information of at least one of the surgery, the procedure, and the research by the medical staff, and the medical staff It may include at least one of performance information and information on whether or not a medical staff has been placed in the hospital.
  • the decision-making of the hospital includes determining the bed assignment of the patient, and the step of performing the decision-making of the hospital may include, based on the medical information about the patient and the predicted patient's medical event, the patient's The bed assignment order may be determined, and the bed assignment of the patient may be determined by further considering the determined bed assignment order of the patient.
  • a computer-readable recording medium is a computer-readable recording medium storing a computer program.
  • the computer program When the computer program is executed by a processor, it stores medical information about a plurality of patients as data in a predetermined structure format. transforming into , and learning a deep learning learning model with at least one medical event that may be generated for the plurality of patients as an input as an input of the data in the predetermined structural format as a correct answer; receiving information and predicting at least one medical event for the patient through the deep learning learning model; It may include instructions for causing the processor to perform a method comprising the step of performing decision-making of the hospital in consideration of the information.
  • the decision-making device of the hospital uses a deep learning learning model to perform medical events (eg, mortality within a predetermined period after patient discharge, readmission rate within a predetermined period after patient discharge, and time of patient readmission) It is possible to predict the incidence rate of heart disease (from the period from to the time of discharge to Therefore, it is possible to provide a patient-specific hospital process.
  • medical events eg, mortality within a predetermined period after patient discharge, readmission rate within a predetermined period after patient discharge, and time of patient readmission
  • FIG. 1 is a block diagram of a decision-making apparatus for a hospital according to an exemplary embodiment.
  • FIG. 2 is a data table for explaining an example of medical treatment information for a patient according to an embodiment.
  • FIG. 3 is a diagram illustrating an example in which medical treatment information for a plurality of patients is converted into data in a predetermined structural format according to an exemplary embodiment.
  • 4A to 4B are diagrams illustrating medical treatment information for a plurality of patients stored in a memory of a decision-making apparatus of a hospital according to an exemplary embodiment.
  • 5A and 5B are diagrams for explaining that data stored in a memory of a decision-making apparatus of a hospital is converted into a standardized structure format and stored according to an exemplary embodiment.
  • FIG. 6 is a flowchart for explaining training a deep learning learning model according to an embodiment.
  • FIG. 7 is a diagram for explaining an example of data input to a deep learning learning model according to an embodiment.
  • FIG. 8 is a flowchart illustrating an example of performing decision-making in a hospital according to an exemplary embodiment.
  • FIG. 9 is a flowchart illustrating an example of performing decision-making in a hospital according to another embodiment.
  • FIG. 10 is an exemplary flowchart of a decision-making method in a hospital according to an embodiment.
  • FIG. 1 is a block diagram of a hospital decision-making apparatus 100 according to an embodiment.
  • a hospital decision-making apparatus 100 may include an input/output unit 101 , a communication unit 102 , a memory 110 , and/or a processor 120 .
  • the input/output unit 101 transmits a command or data input from a user or other external device to other component(s) of the hospital decision-making apparatus 100 according to an embodiment, or in one embodiment A command or data received from other component(s) of the decision-making apparatus 100 of a hospital according to an example may be output to a user or other external device.
  • the input/output unit 101 may receive medical treatment information for a patient.
  • the medical treatment information for the patient may include medical treatment information according to the time when the patient first visits the hospital and medical information according to the time when the patient visits the hospital after the first visit to the hospital. That is, the medical treatment information for the patient may include the patient's past medical treatment information and the patient's current medical information.
  • the medical treatment information for the patient may include treatment information in other hospitals other than the hospital where the patient has currently obtained the medical treatment information.
  • FIG. 2 is a data table for explaining an example of medical treatment information for a patient according to an embodiment.
  • medical treatment information for a patient input to the input/output unit 101 may include hospital visit information (visit table) and medication information (medication table).
  • the hospital visit information includes information about the patient's visit number, visit classification (for example, it can be classified as outpatient, hospitalized, or emergency), hospitalization time and discharge time, or among them (patient's visit number, information on the classification of hospitalization, hospitalization time, and discharge time)) may be included.
  • Medication information includes information on the patient's visit number, drug prescription date, drug type, and drug dose, or information about the patient's visit number, drug prescription date, drug type, and drug dose. ) may include some.
  • the communication unit 102 may support establishment of a wired or wireless communication channel between the hospital decision-making apparatus 100 and an external device, and communication through the established communication channel.
  • Memory 110 is a variety of data used by at least one component (processor 120, input/output unit 101 and/or communication unit 102) of the hospital decision-making apparatus 100, for example, software It can store input data or output data for (eg, program) and related commands.
  • the memory 110 may include a volatile memory or a non-volatile memory.
  • the memory 110 may store medical treatment information for a plurality of patients converted into data of a predetermined structure format.
  • the memory 110 may store a pre-learned deep learning learning model.
  • the medical treatment information for the plurality of patients may be accumulated data of the medical treatment information for the patient input from the input/output unit 101 .
  • treatment information for a plurality of patients stored in the memory 110 will be described with reference to FIGS. 2 to 5 , which is converted into data of a predetermined structural format and stored.
  • FIG. 3 is a diagram illustrating an example in which medical treatment information for a plurality of patients is converted into data in a predetermined structural format according to an exemplary embodiment.
  • the hospital decision-making apparatus 100 may convert medical treatment information for a plurality of patients into data of a predetermined (or standardized) structure format and store the converted data in the memory 110 .
  • text data included in medical treatment information for a plurality of patients may be standardized into numeric data.
  • the division information divided into “outpatient”, “hospitalization” and “emergency” can be converted into “373864002", “416800000” and “4525004", respectively, and the drug information is "stain", Divide by "lnsulin” and “aspirin” can be converted to "1158753”, "106892” and "315431", respectively.
  • the treatment information for a plurality of patients may be converted into data in a standardized structure format using a linkage database, but is not limited thereto.
  • 4A to 4B are diagrams illustrating medical treatment information for a plurality of patients stored in the memory 110 of the hospital decision-making apparatus 100 according to an exemplary embodiment.
  • information about a plurality of patients stored in the memory 110 includes visit time points (Visit 1 to Visit 5) and treatment categories (eg, "hospitalization”, “test”, “medication” and " The data may be sorted according to "operation”).
  • the hospital decision-making apparatus 100 converts the stored data (medical information about a plurality of patients) to other data (visit 1 to visit 5) adjacent to the visit time point (visit 1 to visit 5). It is possible to embed a correlation with medical information for a plurality of patients). In this case, embedding the correlation between the time of each data and other data adjacent in this way is called history-aware embedding.
  • 5A and 5B are diagrams for explaining that data stored in the memory 110 of the hospital decision-making apparatus 100 is converted into a standardized structural format and stored according to an exemplary embodiment.
  • information about a plurality of patients stored in the memory 110 includes visit time points (Visit 1 to Visit 5) and treatment category (eg, “hospitalization”, “examination”, “medication”). and “surgery”) may be sorted according to data, and information on a plurality of patients may be stored as " ⁇ ".
  • information on a plurality of patients when information on a plurality of patients is converted into a standardized structure format and stored, as shown in FIG. 5B , information on a plurality of patients may be changed from “ ⁇ ” to “ ⁇ ”.
  • the processor 120 (also referred to as a control unit, a control device, or a control circuit) includes at least one other component (eg, a hardware component (eg, an input/output unit 101 , a communication unit 102 ) of the connected hospital decision-making device 100 . ) and/or memory 110) or software components), and may perform various data processing and operations.
  • a hardware component eg, an input/output unit 101 , a communication unit 102
  • the processor 120 also referred to as a control unit, a control device, or a control circuit
  • the processor 120 includes at least one other component (eg, a hardware component (eg, an input/output unit 101 , a communication unit 102 ) of the connected hospital decision-making device 100 . ) and/or memory 110) or software components), and may perform various data processing and operations.
  • the processor 120 receives information stored in the memory 110 from the memory 110 (data in which medical information about a plurality of patients is converted into a standardized structural format) as an input, and provides the information to a plurality of patients. After loading a deep learning learning model that has been learned by using at least one medical event that may occur for a patient as a correct answer, it is possible to predict at least one medical event for a patient through the deep learning learning model.
  • the processor 120 may predict at least one medical event for the patient through the deep learning learning model.
  • the processor 120 may be generated for a plurality of patients by inputting information stored in the memory 110 (data in which medical information about a plurality of patients is converted into a standardized structure format) as an input. At least one medical event for the patient may be predicted through the deep learning learning model, which is learned by taking at least one medical event as the correct answer.
  • FIG. 6 is a flowchart for explaining training a deep learning learning model according to an embodiment.
  • the processor 120 receives medical information about a plurality of patients as an input, and at least one medical event that may be generated for the plurality of patients as a correct answer.
  • a deep learning learning model may be trained (step S1).
  • the processor 120 may input some of the medical information for a plurality of patients to the deep learning learning model as input, and check whether a desired medical event is output (step S2). ).
  • the processor 120 may check whether the error value of the medical event output from the deep learning learning model is equal to or less than a preset value (step S3), and the error value of the medical event output from the deep learning learning model is a preset value. is exceeded, the deep learning learning model may be trained again in step S2.
  • the processor 120 when the error value of the medical event output from the deep learning learning model is less than or equal to a preset value, the deep learning learning model may be stored in the memory 110 or other external device (step S4) .
  • the processor 120 may check whether new data (eg, patient's medical information) is introduced (step S5), and in this case, when new data (eg, patient's medical information) is introduced, step Go to S1, you can input new data as input data for training the deep learning learning model. However, it may be terminated if no new data is introduced.
  • new data eg, patient's medical information
  • treatment information for a plurality of patients may be input, and in this case, the treatment information for the plurality of patients is the hospital in which the patient has acquired the current treatment information. It may include medical information from other hospitals other than (Institution A) (Institution B).
  • the processor 120 may predict at least one medical event for the patient by inputting the patient's medical treatment information into the deep learning learning model.
  • the medical event includes the mortality rate within a predetermined period after the patient's discharge, the readmission rate within a predetermined period (for example, may be 30 days) after the patient's discharge, the period from the time the patient is re-admitted to the time of discharge, and the discharge of the patient. It may include at least one of the incidence rates of heart disease within a predetermined period (eg, may be 30 days) after.
  • the mortality rate within a predetermined period after discharge of a patient may include in-hospital death and death due to cancer.
  • the mortality rate within a predetermined period after the patient's discharge was the total number of patients in the hospital (total number of hospitalizations) and the number of patients who died within 30 days from the time of discharge (30 at the time of discharge) among all patients in the hospital. number of deaths within one day).
  • AUROC is a value indicating the accuracy of mortality within a predetermined period after discharge of a patient predicted through a deep learning learning model.
  • the accuracy of mortality within a predetermined period after discharge of a patient predicted through a deep learning learning model is 92.39% by multiplying the value of AUROC by 100.
  • the readmission rate within a predetermined period (for example, may be 30 days) after a patient's discharge is information about readmission within 30 days from the patient's discharge in consideration of the patient's medical information before and during hospitalization.
  • readmission refers to a case where the time of re-admission from the time of discharge is within 30 days only for patients who have been admitted to the hospital at least once.
  • the rate of long-term hospitalized patients among patients re-hospitalized within a predetermined period (for example, may be 30 days) after the patient's discharge was within 30 days among all patients in the hospital.
  • a predetermined period for example, may be 30 days
  • the number of patients and the number of patients whose period from readmission to discharge exceeds 7 days (the number of long-term hospitalizations during readmission (more than 7 days)) among all patients in the hospital can be calculated by
  • AUROC in Table 2 is a value indicating the accuracy of the long-term hospitalization rate among patients re-hospitalized within a predetermined period (for example, it may be 30 days) after discharge of the patient predicted through the deep learning learning model.
  • the accuracy of the long-term hospitalization rate among patients re-hospitalized within a predetermined period (for example, 30 days) after patient discharge through a deep learning learning model is 81.26% by multiplying the value of AUROC by 100 It can be said that
  • Table 3 shows data on patients who were readmitted within 30 days after the patient was discharged, and among them, it is a table indicating whether the period from the time the patient was readmitted to the time of discharge was less than or equal to a predetermined period.
  • the predetermined period is exemplarily expressed as 10 days.
  • the period from the time the patient was re-admitted to the time of discharge is the number of patients who were readmitted within 30 days (the number of readmissions within 30 days) and the total number of patients in the hospital, as shown in Table 3 below. It can be calculated using the number of patients who were readmitted within 30 days and who were hospitalized for less than 10 days (the number of hospitalizations for less than 10 days among them).
  • R ⁇ 2 is an index that can confirm how accurately the number of hospitalization days or hospitalization period was predicted.
  • the mean absolute error may mean the error rate of the period from the time when the predicted patient is re-hospitalized to the time of discharge.
  • the incidence rate of heart disease within a predetermined period (eg, may be 30 days) after a patient's discharge is an event that predicts whether a heart disease will occur within 30 days after discharge among all patients in the hospital.
  • the heart disease includes at least one of death, revascularization, myocardial infarction, and stroke, and the earliest event within 30 days is considered.
  • the incidence rate of heart disease within a predetermined period is the number of patients who visited the hospital again after being discharged from the hospital (the total number of visits) ) and the total number of patients in the hospital with heart disease (eg, death, revascularization, myocardial infarction, stroke) (the number of major cardiac events).
  • the deep learning learning model used to predict medical events is a feed-forward model (FFNN: Feed-forward Neural Network) with boosted time-aware embeddings (or features) and a gradient boosting tree. It may be at least one model among models (GBM: Gradient Boosting Machine).
  • FFNN Feed-forward Neural Network
  • GBM Gradient Boosting Machine
  • the processor 120 performs hospital decision-making in consideration of the predicted patient's medical event, medical information about the patient input from the input/output unit 101, and state information of the hospital (which may be the hospital where the patient has been treated).
  • the hospital status information includes the number of clinics in the hospital, the number of treatment rooms used by each medical staff in the hospital, outpatient support personnel information by medical staff, session opening status information by day of the week for other medical staff in the same department of the hospital, bed information in the hospital, and hospital resources It may include at least one of patient status information, status information for each department, type and number of operating rooms, schedule information for each operating room, and estimated time required for each operation of the patient, but is not limited thereto.
  • the decision-making of the hospital may include at least one of a patient's treatment schedule determination, a patient's operating room schedule determination, and a patient's bed assignment determination.
  • the processor 120 may determine the patient's treatment priority based on the patient's medical event predicted using the patient's medical treatment information and a deep learning learning model.
  • the processor 120 makes a hospital decision in consideration of the patient's medical event predicted using the deep learning learning model, the patient's medical information, the hospital's status information, the patient's medical priority, and the hospital's medical staff information. can be performed.
  • the medical staff information of the hospital includes treatment schedule information by medical staff, information by patient type, time information generated by patient treatment, schedule information of at least one of surgery, procedure and research by medical staff, performance information by medical staff, and medical staff position in the hospital It may include at least one of whether or not information, but is not limited thereto.
  • the processor 120 may determine the patient's bed allocation ranking based on the patient's medical event predicted using the patient's medical treatment information and a deep learning learning model.
  • the processor 120 makes a hospital decision in consideration of the patient's medical event predicted using the deep learning learning model, the medical information about the patient, the state information of the hospital, and the patient's bed (or ward) assignment order can be performed.
  • FIG. 8 is a flowchart illustrating an example of performing decision-making in a hospital according to an exemplary embodiment.
  • the hospital decision-making apparatus 100 may first receive the patient's medical treatment information through the input/output unit 101 (step S11).
  • the hospital decision-making apparatus 100 includes the patient's medical information in the past, including hospitals other than the hospital where the patient obtained the medical information (for example, the patient's hospitalization history record) ) can be checked whether or not there is (step S12).
  • the patient's medical information does not contain the patient's past medical information (for example, the patient's hospitalization history record) including hospitals other than the hospital from which the patient obtained the medical information, the patient's medical priority cannot be maintained. have.
  • the patient's past medical information for example, the patient's hospitalization history record
  • the initial treatment order of the patient may be set as the last priority among the treatment orders of all patients in the hospital, but is not limited thereto.
  • the hospital decision-making device 100 includes past patient medical information (eg, patient's hospitalization history record) including hospitals other than the hospital where the patient obtained medical information in the patient's medical information. It is possible to predict the mortality rate within a predetermined period after a patient's discharge during a medical event through a deep learning learning model on the patient's medical information, and check whether the predicted mortality rate within a predetermined period after discharge exceeds a preset probability (step S13) ).
  • patient medical information eg, patient's hospitalization history record
  • the patient's treatment order may be maintained as it is.
  • the patient's medical treatment order may be increased (step S14).
  • the hospital decision-making apparatus 100 provides the predicted patient's medical event, the patient's medical information, the hospital's state information, the patient's medical priority, and the hospital's medical staff information. Taking this into account, it is possible to determine the patient's treatment schedule during the decision-making of the hospital (step S15).
  • the treatment information about the patient includes the mortality rate within a predetermined period after the patient's discharge, the number of hospital visits before the reference point of the hospital visit, the waiting time for treatment, and the time required for treatment.
  • diagnostic information, medication history information, surgery/procedure/treatment information may include at least one information of hospital visit information including at least one of body measurement information.
  • the medical staff information may be the treatment schedule (for example, the number of openings per week, days of the week, morning and afternoon, total treatment hours, number of patients per time, outpatient return rate (discontinuance rate, additional treatment rate) information), by patient type Information (for example, patient type ratio, capacity, re-treatment rate, reservation rate, actual treatment rate compared to reservation, non-visiting rate, ratio of critically ill patients may be information), average waiting time for each patient, patient's treatment time, It may include information about at least one of patient termination delay time information, medical staff surgery/procedure/research schedule, number of surgical/procedure/research results, post-outpatient surgery linkage rate, and position status information.
  • the treatment schedule for example, the number of openings per week, days of the week, morning and afternoon, total treatment hours, number of patients per time, outpatient return rate (discontinuance rate, additional treatment rate) information
  • patient type Information for example, patient type ratio, capacity, re-treatment rate, reservation rate, actual treatment rate compared to reservation
  • the hospital status information may include at least one of the number of clinics in each department, the number of offices used by medical staff/department, outpatient support personnel for each medical team, and session opening status information by day of the week for other medical staff in the same department.
  • the patient may receive treatment according to the determined patient's treatment schedule.
  • FIG. 9 is a flowchart illustrating an example of performing decision-making in a hospital according to another embodiment.
  • the hospital decision-making apparatus 100 may receive information about a list of patients reserved for hospitalization among medical information about the patient (step S21).
  • the hospital decision-making apparatus 100 predicts the incidence rate of heart disease within a predetermined period after the patient's discharge during a medical event through a deep learning learning model with respect to the patient's medical information (reserved patient list information), , it is possible to check whether the predicted incidence of heart disease within a predetermined period after discharge exceeds a preset probability (step S22), and when the predicted incidence of heart disease within a predetermined period after discharge of the patient is less than or equal to a preset probability, the patient's The hospital bed (or ward) allocation order may be maintained as it is.
  • the initial bed (or ward) assignment order of the patient may be set to the last rank among the bed (or room) assignment orders of all patients in the hospital, but is not limited thereto.
  • the patient's bed (or ward) assignment order may rise (step S23).
  • the hospital decision-making apparatus 100 may provide a patient's medical event predicted according to the patient's bed (or room) assignment order, medical treatment information for the patient, hospital status information, and the patient's bed (or The patient's bed assignment may be determined during decision-making in the hospital by considering the ward) assignment order (step S24).
  • the medical information about the patient includes the patient's registration number (or visit number), name, gender, age, patient type (new/secondary/revisit), etc.
  • Basic information including, diagnosis information, including name of diagnosis, department, doctor in charge, severity level, precautions, special notes, patient status information including type of isolation/receipt date/required period, operation/surgery name, scheduled surgery date, At least one of information about the patient's surgery including the surgeon and the type of anesthesia, the medicine within the hospital stay, the patient's treatment information including the type/prescription/treatment of materials, and the result information of various diagnostic tests and imaging tests. can
  • hospital status information includes total bed information including ward, ward, name/number/number of beds, etc., the number of beds used by all hospitalized patients per ward/ward/bed, At least among inpatient patient status (hospitalization period, surgery/procedure/treatment status, etc.), ward/ward/bed status by department, type and number of operating rooms, schedule status by operating room (surgery name, surgery timetable), and scheduled operation time for each surgery It may contain one piece of information.
  • the total number of patients with a predicted hospitalization period of 31 days or more at the time of admission (the total number of patients with a long hospitalization period), the predicted value of the hospitalization period at the time of admission Predetermined value less than 30 days
  • Total number of patients expected to be hospitalized within N days (eg, N is 3 days) (total number of short-stay patients) and total patients likely to be readmitted within 30 days
  • At least one piece of information among the number of pieces of information may be considered, but is not limited thereto.
  • the patient's bed assignment may succeed or fail, and when the patient's bed assignment succeeds, the patient may perform a hospitalization procedure according to the determined bed assignment.
  • FIG. 10 is an exemplary flowchart of a decision-making method in a hospital according to an embodiment.
  • the hospital decision-making method shown in FIG. 10 can be performed by the hospital decision-making device 100 shown in FIG. 1 .
  • the hospital decision-making method shown in FIG. 10 is merely exemplary.
  • the hospital decision-making apparatus 100 may convert medical treatment information for a plurality of patients into data having a predetermined structure (step S100).
  • the hospital decision-making apparatus 100 may train the deep learning learning model by using data of a predetermined structural format as an input, and at least one medical event that may be generated for a plurality of patients as a correct answer (step S200).
  • the hospital decision-making apparatus 100 may receive medical treatment information for the patient and predict at least one medical event for the patient through the deep learning learning model (step S300 ).
  • the hospital decision-making apparatus 100 may perform the hospital decision-making in consideration of the predicted at least one medical event for the patient, medical treatment information for the patient, and state information of the hospital (step S400 ).
  • the hospital decision-making device uses a deep learning learning model to perform medical events (eg, mortality within a predetermined period after a patient's discharge, a readmission rate within a predetermined period after a patient's discharge, (the period from the patient's re-hospitalization to the time of discharge and the incidence rate of heart disease within a predetermined period after the patient's discharge) Because the system can be managed, it is possible to provide patient-specific hospital processes.
  • medical events eg, mortality within a predetermined period after a patient's discharge, a readmission rate within a predetermined period after a patient's discharge, (the period from the patient's re-hospitalization to the time of discharge and the incidence rate of heart disease within a predetermined period after the patient's discharge).
  • Combinations of each block in the block diagram attached to the present invention and each step in the flowchart may be performed by computer program instructions.
  • These computer program instructions may be embodied in the encoding processor of a general purpose computer, special purpose computer, or other programmable data processing equipment, such that the instructions executed by the encoding processor of the computer or other programmable data processing equipment may correspond to each block of the block diagram or
  • Each step of the flowchart creates a means for performing the functions described.
  • These computer program instructions may also be stored in a computer-usable or computer-readable memory which may direct a computer or other programmable data processing equipment to implement a function in a particular way, and thus the computer-usable or computer-readable memory.
  • the instructions stored in the block diagram may also produce an item of manufacture containing instruction means for performing a function described in each block of the block diagram or each step of the flowchart.
  • the computer program instructions may also be mounted on a computer or other programmable data processing equipment, such that a series of operational steps are performed on the computer or other programmable data processing equipment to create a computer-executed process to create a computer or other programmable data processing equipment. It is also possible that instructions for performing the processing equipment provide steps for carrying out the functions described in each block of the block diagram and in each step of the flowchart.
  • each block or each step may represent a module, segment, or portion of code comprising one or more executable instructions for executing specified logical function(s). It should also be noted that in some alternative embodiments it is also possible for the functions recited in blocks or steps to occur out of order. For example, it is possible that two blocks or steps shown one after another may in fact be performed substantially simultaneously, or that the blocks or steps may sometimes be performed in the reverse order according to the corresponding function.

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • General Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Medical Informatics (AREA)
  • Health & Medical Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Public Health (AREA)
  • Marketing (AREA)
  • Quality & Reliability (AREA)
  • Operations Research (AREA)
  • Tourism & Hospitality (AREA)
  • General Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Primary Health Care (AREA)
  • Epidemiology (AREA)
  • Game Theory and Decision Science (AREA)
  • Development Economics (AREA)
  • Educational Administration (AREA)
  • Software Systems (AREA)
  • Biomedical Technology (AREA)
  • Pathology (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Computation (AREA)
  • Databases & Information Systems (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Medical Treatment And Welfare Office Work (AREA)

Abstract

The present invention relates to an apparatus for decision-making in a hospital. More specifically, the present invention may: convert medical information for a plurality of patients into data in a predetermined structural format; train a deep learning model while marking, as a correct answer, at least one medical event that may occur for the plurality of patients, by inputting the data in the predetermined structural format; receive medical information about a patient and predict at least one medical event for the patient through the deep learning model; and make a decision in a hospital in consideration of the predicted at least one medical event for the patient, the medical information about the patient, and state information of the hospital.

Description

병원의 의사결정 장치, 방법 및 컴퓨터 판독 가능한 기록매체Hospital decision-making apparatus, method, and computer-readable recording medium
본 발명은 병원의 의사결정 장치, 방법, 컴퓨터 판독 가능한 기록매체 및 컴퓨터 프로그램에 관한 것이다.The present invention relates to a hospital decision-making apparatus, method, computer-readable recording medium and computer program.
최근, 병원에서의 재진 적체현상으로 인하여 환자 대기시간이 늘어남에 따라, 환자들의 불편함이 증가하고 있다.Recently, as patient waiting time is increased due to re-diagnostic congestion in hospitals, patient discomfort is increasing.
또한, 병원 내의 복수의 환자들이 진료를 접수한 시간 순서대로 진료, 입원 및 수술 등의 스케줄링을 수행할 경우, 환자 별 특성에 따라 특정 환자의 진료를 우선으로 해야 되는 경우를 고려할 수 없는 문제가 발생할 수 있었다.In addition, when scheduling, such as treatment, hospitalization, and surgery, is performed in the order in which a plurality of patients in the hospital received treatment, a case in which the treatment of a specific patient must be prioritized according to the characteristics of each patient may not be considered. could
따라서, 환자 별 특성을 고려하면서 환자들의 대기 시간은 최소화하고, 병원의 의료자원을 효율적으로 운영하도록 병원의 의사결정을 수행할 수 있는 기술을 필요로 하는 실정이다.Therefore, while considering the characteristics of each patient, the patient's waiting time is minimized, and there is a need for a technology capable of performing hospital decision-making to efficiently operate the hospital's medical resources.
본 발명의 해결하고자 하는 과제는, 병원의 의사결정 장치, 방법, 컴퓨터 판독 가능한 기록매체 및 컴퓨터 프로그램을 제공하는 것이다.An object of the present invention is to provide a hospital decision-making apparatus, a method, a computer-readable recording medium, and a computer program.
또한, 이러한 병원의 의사결정 장치는 딥러닝 학습모델을 통해 환자의 의료 이벤트를 예측하고, 예측된 의료 이벤트, 환자에 대한 진료정보 및 병원의 상태정보를 종합적으로 고려하여 병원의 의사결정을 수행하는 것 등이 본 발명의 해결하고자 하는 과제에 포함될 수 있다.In addition, this hospital decision-making device predicts a patient's medical event through a deep learning learning model, and comprehensively considers the predicted medical event, the patient's medical information, and the hospital's state information to make the hospital's decision. and the like may be included in the problems to be solved by the present invention.
다만, 본 발명의 해결하고자 하는 과제는 이상에서 언급한 것으로 제한되지 않으며, 언급되지 않은 또 다른 해결하고자 하는 과제는 아래의 기재로부터 본 발명이 속하는 통상의 지식을 가진 자에게 명확하게 이해될 수 있을 것이다.However, the problems to be solved of the present invention are not limited to those mentioned above, and other problems to be solved that are not mentioned can be clearly understood by those of ordinary skill in the art to which the present invention belongs from the following description. will be.
일 실시예에 따른 병원의 의사결정 장치는, 환자에 대한 진료정보를 입력 받는 입출력부; 복수의 환자에 대한 진료정보가 기 정해진 구조 형식의 데이터로 변환되어 저장되어 있는 메모리; 및 상기 메모리와 전기적으로 연결되는 프로세서를 포함하며, 상기 프로세서는, 상기 기 정해진 구조 형식의 데이터를 입력으로, 상기 복수의 환자에 대해 발생될 수 있는 적어도 하나의 의료 이벤트를 정답으로하여 학습된, 딥러닝 학습모델을 통해 상기 환자에 대한 적어도 하나의 의료 이벤트를 예측하고, 상기 예측된 상기 환자에 대한 적어도 하나의 의료 이벤트, 상기 환자에 대한 진료정보 및 병원의 상태정보를 고려하여 상기 병원의 의사결정을 수행할 수 있다.A hospital decision-making apparatus according to an embodiment includes an input/output unit for receiving medical treatment information about a patient; a memory in which medical information for a plurality of patients is converted into data of a predetermined structure format and stored; and a processor electrically connected to the memory, wherein the processor receives the data in the predetermined structural format as an input, and learns by answering at least one medical event that may occur for the plurality of patients as an answer. A doctor of the hospital predicts at least one medical event for the patient through a deep learning learning model, and considers the predicted at least one medical event for the patient, medical treatment information for the patient, and state information of the hospital decision can be made.
또한, 상기 환자에 대한 진료정보는, 상기 환자가 상기 병원을 최초 방문한 시점에 따른 진료정보, 상기 환자가 상기 병원을 최초 방문한 시점 이후에 상기 병원을 방문한 시점에 따른 진료정보 및 상기 병원 이외의 타 병원에서 획득한 진료정보 중 적어도 하나를 포함할 수 있다.In addition, the medical treatment information for the patient includes medical treatment information according to the time when the patient first visits the hospital, medical information according to the time when the patient visits the hospital after the initial visit to the hospital, and other information other than the hospital It may include at least one of the medical information obtained from the hospital.
또한, 상기 의료 이벤트는, 상기 환자의 퇴원 후 소정기간 이내 사망률, 상기 환자의 퇴원 후 소정기간 이내 재입원률, 상기 환자가 재입원한 시점에서부터 퇴원시점까지의 기간 및 상기 환자의 퇴원 후 소정기간 이내 심장병의 발병률 중 적어도 하나를 포함할 수 있다.In addition, the medical event may include a mortality rate within a predetermined period after the patient's discharge, a readmission rate within a predetermined period after the patient's discharge, a period from the time the patient reads into the hospital to a discharge time, and within a predetermined period after the patient's discharge at least one of the incidence of heart disease.
또한, 상기 병원의 의사결정은, 상기 환자의 진료 스케줄 결정, 상기 환자의 병상배정 결정 및 상기 환자의 수술실 스케줄 결정을 포함할 수 있다.In addition, the decision-making of the hospital may include determining the patient's treatment schedule, determining the bed assignment of the patient, and determining the operating room schedule of the patient.
또한, 상기 병원의 상태정보는, 상기 병원의 진료실 개수, 상기 병원의 의료진 별 사용한 진료실의 개수, 상기 의료진 별 외래 지원 인력정보, 상기 병원의 동일 진료과 내 타 의료진 요일별 세션 개설 상태정보, 상기 병원의 병실 침대정보, 상기 병원의 재원 환자 상태정보, 상기 진료과 별 상태정보, 상기 병원의 수술실의 종류와 개수, 상기 수술실별 스케줄 정보 및 상기 환자의 수술별 소요예정시간 정보 중 적어도 하나를 포함할 수 있다.In addition, the state information of the hospital includes the number of treatment rooms in the hospital, the number of treatment rooms used by each medical staff of the hospital, outpatient support personnel information for each medical staff, session opening status information for each day of the other medical staff in the same department of the hospital, and the hospital's It may include at least one of bed information in a hospital room, patient status information of the hospital, status information for each department, the type and number of operating rooms in the hospital, schedule information for each operating room, and estimated time required for each operation of the patient. .
또한, 상기 프로세서는, 상기 병원의 의사결정은, 상기 환자의 진료 스케줄 결정을 포함하고, 상기 환자에 대한 진료정보 및 상기 예측된 환자의 의료 이벤트를 기초로 상기 환자의 진료 순위를 결정하고, 상기 결정된 환자의 진료 순위 및 상기 병원의 의료진 정보를 더 고려하여 상기 환자의 진료 스케줄을 결정할 수 있다.In addition, the processor, the decision-making of the hospital includes determining the patient's medical schedule, and determining the patient's medical care ranking based on the medical treatment information for the patient and the predicted patient's medical event, The patient's treatment schedule may be determined by further considering the determined patient's treatment order and medical staff information of the hospital.
또한, 상기 병원의 의료진 정보는, 상기 의료진 별 진료 스케줄정보, 상기 환자 유형별 정보, 상기 환자의 진료에 의해 발생되는 시간정보, 상기 의료진 별 수술, 시술 및 연구 중 적어도 하나의 스케줄 정보, 상기 의료진 별 실적정보 및 상기 병원의 의료진 보직 여부 정보 중 적어도 하나를 포함할 수 있다.In addition, the medical staff information of the hospital includes the medical treatment schedule information for each medical staff, the information by the patient type, the time information generated by the patient's treatment, the schedule information of at least one of the surgery, the procedure, and the research by the medical staff, and the medical staff It may include at least one of performance information and information on whether or not a medical staff has been placed in the hospital.
또한, 상기 병원의 의사결정은, 상기 환자의 병상배정 결정을 포함하고, 상기 프로세서는, 상기 환자에 대한 진료정보 및 상기 예측된 환자의 의료 이벤트를 기초로 상기 환자의 병상배정 순위를 결정하고, 상기 결정된 환자의 병상 배정 순위를 더 고려하여 상기 환자의 병상배정을 결정할 수 있다.In addition, the decision-making of the hospital includes determining the bed assignment of the patient, and the processor determines the bed assignment rank of the patient based on the medical event information of the patient and the predicted patient, The bed assignment of the patient may be determined by further considering the determined bed assignment order of the patient.
일 실시예에 따른 병원의 의사결정 방법은, 복수의 환자에 대한 진료정보를 기 정해진 구조 형식의 데이터로 변환시키는 단계와, 상기 기 정해진 구조 형식의 데이터를 입력으로, 상기 복수의 환자에 대해 발생될 수 있는 적어도 하나의 의료 이벤트를 정답으로하여 딥러닝 학습모델을 학습시키는 단계와, 환자에 대한 진료정보를 입력 받아, 상기 딥러닝 학습모델을 통해 상기 환자에 대한 적어도 하나의 의료 이벤트를 예측하는 단계와, 상기 예측된 상기 환자에 대한 적어도 하나의 의료 이벤트, 상기 환자에 대한 진료정보 및 병원의 상태정보를 고려하여 상기 병원의 의사결정을 수행하는 단계를 포함할 수 있다.A hospital decision-making method according to an embodiment includes converting medical treatment information for a plurality of patients into data of a predetermined structural format, and generating data in the predetermined structural format as an input for the plurality of patients Learning a deep learning learning model with at least one possible medical event as the correct answer, receiving medical treatment information about a patient, and predicting at least one medical event for the patient through the deep learning learning model and performing the decision-making of the hospital in consideration of the predicted at least one medical event for the patient, medical treatment information for the patient, and state information of the hospital.
또한, 상기 환자에 대한 진료정보는, 상기 환자가 상기 병원을 최초 방문한 시점에 따른 진료정보, 상기 환자가 상기 병원을 최초 방문한 시점 이후에 상기 병원을 방문한 시점에 따른 진료정보 및 상기 병원 이외의 타 병원에서 획득한 진료정보 중 적어도 하나를 포함할 수 있다.In addition, the medical treatment information for the patient includes medical treatment information according to the time when the patient first visits the hospital, medical information according to the time when the patient visits the hospital after the initial visit to the hospital, and other information other than the hospital It may include at least one of the medical information obtained from the hospital.
또한, 상기 의료 이벤트는, 상기 환자의 퇴원 후 소정기간 이내 사망률, 상기 환자의 퇴원 후 소정기간 이내 재입원률, 상기 환자가 재입원한 시점에서부터 퇴원시점까지의 기간 및 상기 환자의 퇴원 후 소정기간 이내 심장병의 발병률 중 적어도 하나를 포함할 수 있다.In addition, the medical event may include a mortality rate within a predetermined period after the patient's discharge, a readmission rate within a predetermined period after the patient's discharge, a period from the time the patient reads into the hospital to a discharge time, and within a predetermined period after the patient's discharge at least one of the incidence of heart disease.
또한, 상기 병원의 의사결정은, 상기 환자의 진료 스케줄 결정, 상기 환자의 병상배정 결정 및 상기 환자의 수술실 스케줄 결정을 포함할 수 있다.In addition, the decision-making of the hospital may include determining the patient's treatment schedule, determining the bed assignment of the patient, and determining the operating room schedule of the patient.
또한, 상기 병원의 상태정보는, 상기 병원의 진료실 개수, 상기 병원의 의료진 별 사용한 진료실의 개수, 상기 의료진 별 외래 지원 인력정보, 상기 병원의 동일 진료과 내 타 의료진 요일별 세션 개설 상태정보, 상기 병원의 병실 침대정보, 상기 병원의 재원 환자 상태정보, 상기 진료과 별 상태정보, 상기 병원의 수술실의 종류와 개수, 상기 수술실별 스케줄 정보 및 상기 환자의 수술별 소요예정시간 정보 중 적어도 하나를 포함할 수 있다.In addition, the state information of the hospital includes the number of treatment rooms in the hospital, the number of treatment rooms used by each medical staff of the hospital, outpatient support personnel information for each medical staff, session opening status information for each day of the other medical staff in the same department of the hospital, and the hospital's It may include at least one of bed information in a hospital room, patient status information of the hospital, status information for each department, the type and number of operating rooms in the hospital, schedule information for each operating room, and estimated time required for each operation of the patient. .
또한, 상기 병원의 의사결정은, 상기 환자의 진료 스케줄 결정을 포함하고, 상기 병원의 의사결정을 수행하는 단계는, 상기 환자에 대한 진료정보 및 상기 예측된 환자의 의료 이벤트를 기초로 상기 환자의 진료 순위를 결정하고, 상기 결정된 환자의 진료 순위 및 상기 병원의 의료진 정보를 더 고려하여 상기 환자의 진료 스케줄을 결정할 수 있다.In addition, the decision-making of the hospital includes determining the treatment schedule of the patient, and the performing of the decision-making of the hospital includes the treatment of the patient based on the treatment information about the patient and the predicted medical event of the patient. The treatment order may be determined, and the treatment schedule of the patient may be determined by further considering the determined treatment order of the patient and the medical staff information of the hospital.
또한, 상기 병원의 의료진 정보는, 상기 의료진 별 진료 스케줄정보, 상기 환자 유형별 정보, 상기 환자의 진료에 의해 발생되는 시간정보, 상기 의료진 별 수술, 시술 및 연구 중 적어도 하나의 스케줄 정보, 상기 의료진 별 실적정보 및 상기 병원의 의료진 보직 여부 정보 중 적어도 하나를 포함할 수 있다.In addition, the medical staff information of the hospital includes the medical treatment schedule information for each medical staff, the information by the patient type, the time information generated by the patient's treatment, the schedule information of at least one of the surgery, the procedure, and the research by the medical staff, and the medical staff It may include at least one of performance information and information on whether or not a medical staff has been placed in the hospital.
또한, 상기 병원의 의사결정은, 상기 환자의 병상배정 결정을 포함하고, 상기 병원의 의사결정을 수행하는 단계는, 상기 환자에 대한 진료정보 및 상기 예측된 환자의 의료 이벤트를 기초로 상기 환자의 병상배정 순위를 결정하고, 상기 결정된 환자의 병상 배정 순위를 더 고려하여 상기 환자의 병상배정을 결정할 수 있다.In addition, the decision-making of the hospital includes determining the bed assignment of the patient, and the step of performing the decision-making of the hospital may include, based on the medical information about the patient and the predicted patient's medical event, the patient's The bed assignment order may be determined, and the bed assignment of the patient may be determined by further considering the determined bed assignment order of the patient.
일 실시예에 따른 컴퓨터 판독 가능한 기록매체는, 컴퓨터 프로그램을 저장하고 있는 컴퓨터 판독 가능 기록매체로서, 상기 컴퓨터 프로그램은, 프로세서에 의해 실행되면, 복수의 환자에 대한 진료정보를 기 정해진 구조 형식의 데이터로 변환시키는 단계와, 상기 기 정해진 구조 형식의 데이터를 입력으로, 상기 복수의 환자에 대해 발생될 수 있는 적어도 하나의 의료 이벤트를 정답으로하여 딥러닝 학습모델을 학습시키는 단계와, 환자에 대한 진료정보를 입력 받아, 상기 딥러닝 학습모델을 통해 상기 환자에 대한 적어도 하나의 의료 이벤트를 예측하는 단계와, 상기 예측된 상기 환자에 대한 적어도 하나의 의료 이벤트, 상기 환자에 대한 진료정보 및 병원의 상태정보를 고려하여 상기 병원의 의사결정을 수행하는 단계를 포함하는 방법을 상기 프로세서가 수행하도록 하기 위한 명령어를 포함할 수 있다.A computer-readable recording medium according to an embodiment is a computer-readable recording medium storing a computer program. When the computer program is executed by a processor, it stores medical information about a plurality of patients as data in a predetermined structure format. transforming into , and learning a deep learning learning model with at least one medical event that may be generated for the plurality of patients as an input as an input of the data in the predetermined structural format as a correct answer; receiving information and predicting at least one medical event for the patient through the deep learning learning model; It may include instructions for causing the processor to perform a method comprising the step of performing decision-making of the hospital in consideration of the information.
일 실시예에 따르면 병원의 의사결정 장치는 딥러닝 학습 모델을 이용하여 의료 이벤트(예를 들어, 환자의 퇴원 후 소정기간 이내 사망률, 환자의 퇴원 후 소정기간 이내 재입원률, 환자가 재입원한 시점에서부터 퇴원시점까지의 기간 및 환자의 퇴원 후 소정기간 이내 심장병의 발병률)를 예측하고, 예측된 의료 이벤트, 환자에 대한 진료정보 및 병원의 상태정보를 종합적으로 고려하여 병원 내의 시스템을 관리할 수 있기 때문에 환자 맞춤형 병원 프로세스를 제공할 수 있다.According to an embodiment, the decision-making device of the hospital uses a deep learning learning model to perform medical events (eg, mortality within a predetermined period after patient discharge, readmission rate within a predetermined period after patient discharge, and time of patient readmission) It is possible to predict the incidence rate of heart disease (from the period from to the time of discharge to Therefore, it is possible to provide a patient-specific hospital process.
도 1은 일 실시예에 따른 병원의 의사결정 장치의 블록도이다.1 is a block diagram of a decision-making apparatus for a hospital according to an exemplary embodiment.
도 2는 일 실시예에 따른 환자에 대한 진료정보의 예시를 설명하기 위한 데이터 테이블이다.2 is a data table for explaining an example of medical treatment information for a patient according to an embodiment.
도 3은 일 실시예에 따른 복수의 환자에 대한 진료정보를 기 정해진 구조 형식의 데이터로 변환한 예를 나타낸 도면이다.3 is a diagram illustrating an example in which medical treatment information for a plurality of patients is converted into data in a predetermined structural format according to an exemplary embodiment.
도 4a 내지 도 4b는 일 실시예에 따른 병원의 의사결정 장치의 메모리에 저장되는 복수의 환자에 대한 진료정보를 나타낸 도면이다.4A to 4B are diagrams illustrating medical treatment information for a plurality of patients stored in a memory of a decision-making apparatus of a hospital according to an exemplary embodiment.
도 5a 및 도 5b는 일 실시예에 따른 병원의 의사결정 장치의 메모리에 저장되는 데이터가 표준화된 구조 형식으로 변환되어 저장되는 것을 설명하기 위한 도면이다.5A and 5B are diagrams for explaining that data stored in a memory of a decision-making apparatus of a hospital is converted into a standardized structure format and stored according to an exemplary embodiment.
도 6은 일 실시예에 따른 딥러닝 학습모델을 학습시키는 것을 설명하기 위한 흐름도이다.6 is a flowchart for explaining training a deep learning learning model according to an embodiment.
도 7은 일 실시예에 따른 딥러닝 학습모델에 입력되는 데이터의 예를 설명하기 위한 도면이다.7 is a diagram for explaining an example of data input to a deep learning learning model according to an embodiment.
도 8은 일 실시예에 따른 병원의 의사결정을 수행하는 예를 설명하기 위한 흐름도이다.8 is a flowchart illustrating an example of performing decision-making in a hospital according to an exemplary embodiment.
도 9는 또 다른 실시예에 따른 병원의 의사결정을 수행하는 예를 설명하기 위한 흐름도이다.9 is a flowchart illustrating an example of performing decision-making in a hospital according to another embodiment.
도 10은 일 실시예에 따른 병원의 의사결정 방법에 대한 예시적인 흐름도이다.10 is an exemplary flowchart of a decision-making method in a hospital according to an embodiment.
본 발명의 이점 및 특징, 그리고 그것들을 달성하는 방법은 첨부되는 도면과 함께 상세하게 후술되어 있는 실시예들을 참조하면 명확해질 것이다. 그러나 본 발명은 이하에서 개시되는 실시예들에 한정되는 것이 아니라 서로 다른 다양한 형태로 구현될 수 있으며, 단지 본 실시예들은 본 발명의 개시가 완전하도록 하고, 본 발명이 속하는 기술분야에서 통상의 지식을 가진 자에게 발명의 범주를 완전하게 알려주기 위해 제공되는 것이며, 본 발명은 청구항의 범주에 의해 정의될 뿐이다.Advantages and features of the present invention and methods of achieving them will become apparent with reference to the embodiments described below in detail in conjunction with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but may be implemented in various different forms, and only these embodiments allow the disclosure of the present invention to be complete, and common knowledge in the art to which the present invention pertains It is provided to fully inform those who have the scope of the invention, and the present invention is only defined by the scope of the claims.
본 발명의 실시예들을 설명함에 있어서 공지 기능 또는 구성에 대한 구체적인 설명이 본 발명의 요지를 불필요하게 흐릴 수 있다고 판단되는 경우에는 그 상세한 설명을 생략할 것이다. 그리고 후술되는 용어들은 본 발명의 실시예에서의 기능을 고려하여 정의된 용어들로서 이는 사용자, 운용자의 의도 또는 관례 등에 따라 달라질 수 있다. 그러므로 그 정의는 본 명세서 전반에 걸친 내용을 토대로 내려져야 할 것이다.In describing the embodiments of the present invention, if it is determined that a detailed description of a well-known function or configuration may unnecessarily obscure the gist of the present invention, the detailed description thereof will be omitted. In addition, the terms to be described later are terms defined in consideration of functions in an embodiment of the present invention, which may vary according to intentions or customs of users and operators. Therefore, the definition should be made based on the content throughout this specification.
도 1은 일 실시예에 따른 병원의 의사결정 장치(100)의 블록도이다.1 is a block diagram of a hospital decision-making apparatus 100 according to an embodiment.
도 1을 참조하면, 일 실시예에 따른 병원의 의사결정 장치(100)는 입출력부(101), 통신부(102), 메모리(110) 및/또는 프로세서(120)를 포함할 수 있다.Referring to FIG. 1 , a hospital decision-making apparatus 100 according to an embodiment may include an input/output unit 101 , a communication unit 102 , a memory 110 , and/or a processor 120 .
입출력부(101)는, 예를 들면, 사용자 또는 다른 외부 기기로부터 입력된 명령 또는 데이터를 일 실시예에 따른 병원의 의사결정 장치(100)의 다른 구성요소(들)에 전달하거나, 또는 일 실시예에 따른 병원의 의사결정 장치(100)의 다른 구성요소(들)로부터 수신된 명령 또는 데이터를 사용자 또는 다른 외부 기기로 출력할 수 있다.The input/output unit 101, for example, transmits a command or data input from a user or other external device to other component(s) of the hospital decision-making apparatus 100 according to an embodiment, or in one embodiment A command or data received from other component(s) of the decision-making apparatus 100 of a hospital according to an example may be output to a user or other external device.
일 실시예로서, 입출력부(101)는 환자에 대한 진료정보를 입력 받을 수 있다.As an embodiment, the input/output unit 101 may receive medical treatment information for a patient.
여기서, 환자에 대한 진료정보는 환자가 병원을 최초 방문한 시점에 따른 진료정보와 환자가 병원을 최초 방문한 시점 이후에 병원을 방문한 시점에 따른 진료정보를 포함할 수 있다. 즉, 환자에 대한 진료정보는 환자의 과거 진료정보와 환자의 현재 진료정보를 포함할 수 있다.Here, the medical treatment information for the patient may include medical treatment information according to the time when the patient first visits the hospital and medical information according to the time when the patient visits the hospital after the first visit to the hospital. That is, the medical treatment information for the patient may include the patient's past medical treatment information and the patient's current medical information.
또한, 환자에 대한 진료정보는, 환자가 현재 진료정보를 획득한 병원 이외의 타 병원에서의 진료정보를 포함할 수도 있다.In addition, the medical treatment information for the patient may include treatment information in other hospitals other than the hospital where the patient has currently obtained the medical treatment information.
이하, 입출력부(101)에서 입력 받는 환자에 대한 진료정보에 대한 예시를 도 2를 참조하여 설명하도록 한다.Hereinafter, an example of medical treatment information for a patient input from the input/output unit 101 will be described with reference to FIG. 2 .
도 2는 일 실시예에 따른 환자에 대한 진료정보의 예시를 설명하기 위한 데이터 테이블이다.2 is a data table for explaining an example of medical treatment information for a patient according to an embodiment.
도 2를 참조하면, 입출력부(101)에 입력되는 환자에 대한 진료정보는 병원 방문정보(방문 테이블) 및 투약정보(투약 테이블)를 포함할 수 있다. Referring to FIG. 2 , medical treatment information for a patient input to the input/output unit 101 may include hospital visit information (visit table) and medication information (medication table).
이때, 병원 방문정보는 환자의 내원번호, 내원구분(예를 들어, 외래, 입원, 응급으로 구분될 수 있음), 입원시간 및 퇴원시간에 관한 정보를 포함하거나, 이 중(환자의 내원번호, 내원구분, 입원시간 및 퇴원시간에 관한 정보) 일부를 포함할 수 있다.In this case, the hospital visit information includes information about the patient's visit number, visit classification (for example, it can be classified as outpatient, hospitalized, or emergency), hospitalization time and discharge time, or among them (patient's visit number, information on the classification of hospitalization, hospitalization time, and discharge time)) may be included.
투약정보(투약 테이블)는 환자의 내원번호, 약물 처방날짜, 약물종류 및 약물의 용량에 관한 정보를 포함거나, 이 중(환자의 내원번호, 약물 처방날짜, 약물종류 및 약물의 용량에 관한 정보) 일부를 포함할 수 있다.Medication information (medication table) includes information on the patient's visit number, drug prescription date, drug type, and drug dose, or information about the patient's visit number, drug prescription date, drug type, and drug dose. ) may include some.
통신부(102)는 병원의 의사결정 장치(100)와 외부 장치와의 유선 또는 무선 통신 채널의 수립, 및 수립된 통신 채널을 통한 통신 수행을 지원할 수 있다.The communication unit 102 may support establishment of a wired or wireless communication channel between the hospital decision-making apparatus 100 and an external device, and communication through the established communication channel.
메모리(110)는 병원의 의사결정 장치(100)의 적어도 하나의 구성요소(프로세서(120), 입출력부(101) 및/또는 통신부(102))에 의해 사용되는 다양한 데이터, 예를 들어, 소프트웨어(예: 프로그램) 및, 이와 관련된 명령에 대한 입력 데이터 또는 출력 데이터를 저장할 수 있다. 메모리(110)는, 휘발성 메모리 또는 비휘발성 메모리를 포함할 수 있다. Memory 110 is a variety of data used by at least one component (processor 120, input/output unit 101 and/or communication unit 102) of the hospital decision-making apparatus 100, for example, software It can store input data or output data for (eg, program) and related commands. The memory 110 may include a volatile memory or a non-volatile memory.
일 실시예로서, 메모리(110)는 복수의 환자에 대한 진료정보가 기 정해진 구조 형식의 데이터로 변환되어 저장되어 있을 수 있다.As an embodiment, the memory 110 may store medical treatment information for a plurality of patients converted into data of a predetermined structure format.
예컨대, 메모리(110)는 기 학습된 딥러닝 학습모델이 저장되어 있을 수 있다.For example, the memory 110 may store a pre-learned deep learning learning model.
이때, 복수의 환자에 대한 진료정보는 입출력부(101)에서 입력 받은 환자에 대한 진료정보의 누적 데이터일 수 있다. In this case, the medical treatment information for the plurality of patients may be accumulated data of the medical treatment information for the patient input from the input/output unit 101 .
이하, 도 2 내지 도 5를 통해 메모리(110)에 저장되는 복수의 환자에 대한 진료정보가 기 정해진 구조 형식의 데이터로 변환되어 저장되는 것에 대하여 설명하도록 한다.Hereinafter, treatment information for a plurality of patients stored in the memory 110 will be described with reference to FIGS. 2 to 5 , which is converted into data of a predetermined structural format and stored.
도 3은 일 실시예에 따른 복수의 환자에 대한 진료정보를 기 정해진 구조 형식의 데이터로 변환한 예를 나타낸 도면이다.3 is a diagram illustrating an example in which medical treatment information for a plurality of patients is converted into data in a predetermined structural format according to an exemplary embodiment.
일 실시예에 따른 병원의 의사결정 장치(100)는 복수의 환자에 대한 진료정보를 기 정해진(또는, 표준화된) 구조 형식의 데이터로 변환시켜 메모리(110)에 저장할 수 있다.The hospital decision-making apparatus 100 according to an embodiment may convert medical treatment information for a plurality of patients into data of a predetermined (or standardized) structure format and store the converted data in the memory 110 .
도 3을 참조하면, 복수의 환자에 대한 진료정보에 포함된 문자 데이터가 숫자 데이터로 표준화될 수 있다.Referring to FIG. 3 , text data included in medical treatment information for a plurality of patients may be standardized into numeric data.
예를 들어, 도 2에서 내원구분 정보를 "외래", "입원" 및 "응급"으로 나눈 것을 각각 "373864002", "416800000" 및 "4525004"로 변환시킬 수 있고, 약물정보를 "stain", "lnsulin" 및 "aspirin"으로 나눈 것을 각각 "1158753", "106892" 및 "315431"으로 변환시킬 수 있다.For example, in FIG. 2, the division information divided into "outpatient", "hospitalization" and "emergency" can be converted into "373864002", "416800000" and "4525004", respectively, and the drug information is "stain", Divide by "lnsulin" and "aspirin" can be converted to "1158753", "106892" and "315431", respectively.
이때, 복수의 환자에 대한 진료정보는 linkage database를 이용하여 표준화된 구조 형식의 데이터로 변환될 수 있으나, 이에 한정되는 것은 아니다.In this case, the treatment information for a plurality of patients may be converted into data in a standardized structure format using a linkage database, but is not limited thereto.
도 4a 내지 도 4b는 일 실시예에 따른 병원의 의사결정 장치(100)의 메모리(110)에 저장되는 복수의 환자에 대한 진료정보를 나타낸 도면이다.4A to 4B are diagrams illustrating medical treatment information for a plurality of patients stored in the memory 110 of the hospital decision-making apparatus 100 according to an exemplary embodiment.
도 4a를 참조하면, 메모리(110)에 저장되는 복수의 환자에 대한 정보는 방문 시점(방문 1 내지 방문 5) 및 진료구분(예를 들어,"입원", "검사", "투약" 및 "수술"로 구분될 수 있음)에 따라 데이터가 정렬될 수 있다. Referring to FIG. 4A , information about a plurality of patients stored in the memory 110 includes visit time points (Visit 1 to Visit 5) and treatment categories (eg, "hospitalization", "test", "medication" and " The data may be sorted according to "operation").
이때, 일 실시예에 따른 병원의 의사결정 장치(100)는 도 4b에서 도시된 바와 같이, 저장된 데이터(복수의 환자에 대한 진료정보)를 방문 시점(방문 1 내지 방문 5)이 인접한 다른 데이터(복수의 환자에 대한 진료정보)와의 상관관계를 임베딩할 수 있다. 이때, 이와 같이 각 데이터의 시간이 인접한 다른 데이터와의 상관관계를 임베딩하는 것을 과거 인식(history-aware) 임베딩이라고 한다.At this time, as shown in FIG. 4B , the hospital decision-making apparatus 100 according to an embodiment converts the stored data (medical information about a plurality of patients) to other data (visit 1 to visit 5) adjacent to the visit time point (visit 1 to visit 5). It is possible to embed a correlation with medical information for a plurality of patients). In this case, embedding the correlation between the time of each data and other data adjacent in this way is called history-aware embedding.
도 5a 및 도 5b는 일 실시예에 따른 병원의 의사결정 장치(100)의 메모리(110)에 저장되는 데이터가 표준화된 구조형식으로 변환되어 저장되는 것을 설명하기 위한 도면이다.5A and 5B are diagrams for explaining that data stored in the memory 110 of the hospital decision-making apparatus 100 is converted into a standardized structural format and stored according to an exemplary embodiment.
도 5a 에서 도시된 바와 같이, 메모리(110)에 저장되는 복수의 환자에 대한 정보는 방문 시점(방문 1 내지 방문 5) 및 진료구분(예를 들어,"입원", "검사", "투약" 및 "수술"로 구분될 수 있음)에 따라 데이터가 정렬되며, 복수의 환자에 대한 정보는 "○"로 저장될 수 있다.As shown in FIG. 5A , information about a plurality of patients stored in the memory 110 includes visit time points (Visit 1 to Visit 5) and treatment category (eg, “hospitalization”, “examination”, “medication”). and "surgery") may be sorted according to data, and information on a plurality of patients may be stored as "○".
이때, 복수의 환자에 대한 정보가 표준화된 구조 형식으로 변환되어 저장될 경우, 도 5b에서 도시된 바와 같이, 복수의 환자에 대한 정보는"○"에서 "□"으로 변경될 수 있다.In this case, when information on a plurality of patients is converted into a standardized structure format and stored, as shown in FIG. 5B , information on a plurality of patients may be changed from “○” to “□”.
프로세서(120)(제어부, 제어 장치 또는 제어 회로라고도 함)는 연결된 병원의 의사결정 장치(100)의 적어도 하나의 다른 구성요소(예: 하드웨어 구성 요소(예: 입출력 부(101), 통신부(102) 및/또는 메모리(110)) 또는 소프트웨어 구성요소)를 제어할 수 있고, 다양한 데이터 처리 및 연산을 수행할 수 있다.The processor 120 (also referred to as a control unit, a control device, or a control circuit) includes at least one other component (eg, a hardware component (eg, an input/output unit 101 , a communication unit 102 ) of the connected hospital decision-making device 100 . ) and/or memory 110) or software components), and may perform various data processing and operations.
일 실시예로서, 프로세서(120)는 메모리(110)로부터 메모리(110)에 저장되어 있는 정보(복수의 환자에 대한 진료정보가 표준화된 구조 형식으로 변환된 데이터)를 입력으로, 복수의 환자에 대해 발생될 수 있는 적어도 하나의 의료 이벤트를 정답으로하여 학습된, 딥러닝 학습모델을 로드한 후, 딥러닝 학습모델을 통해 환자에 대한 적어도 하나의 의료 이벤트를 예측할 수 있다.As an embodiment, the processor 120 receives information stored in the memory 110 from the memory 110 (data in which medical information about a plurality of patients is converted into a standardized structural format) as an input, and provides the information to a plurality of patients. After loading a deep learning learning model that has been learned by using at least one medical event that may occur for a patient as a correct answer, it is possible to predict at least one medical event for a patient through the deep learning learning model.
다른 실시예로서, 프로세서(120)는 외부의 다른 장치로부터 딥러닝 학습 모델을 로드한 후, 딥러닝 학습모델을 통해 환자에 대한 적어도 하나의 의료 이벤트를 예측할 수 있다.As another embodiment, after loading a deep learning learning model from another external device, the processor 120 may predict at least one medical event for the patient through the deep learning learning model.
일 실시예로서, 프로세서(120)는 메모리(110)에 저장되어 있는 정보(복수의 환자에 대한 진료정보가 표준화된 구조 형식으로 변환된 데이터)를 입력으로, 복수의 환자에 대해 발생될 수 있는 적어도 하나의 의료 이벤트를 정답으로하여 학습된, 딥러닝 학습모델을 통해 환자에 대한 적어도 하나의 의료 이벤트를 예측할 수 있다.As an embodiment, the processor 120 may be generated for a plurality of patients by inputting information stored in the memory 110 (data in which medical information about a plurality of patients is converted into a standardized structure format) as an input. At least one medical event for the patient may be predicted through the deep learning learning model, which is learned by taking at least one medical event as the correct answer.
이하, 프로세서(120)에서 딥러닝 학습모델을 학습시키는 것에 대하여 도 6 및 도 7을 참조하여 설명하도록 한다.Hereinafter, training of the deep learning learning model in the processor 120 will be described with reference to FIGS. 6 and 7 .
도 6은 일 실시예에 따른 딥러닝 학습모델을 학습시키는 것을 설명하기 위한 흐름도이다.6 is a flowchart for explaining training a deep learning learning model according to an embodiment.
도 6을 참조하면, 먼저 딥러닝 학습모델을 학습하기 위하여 프로세서(120)는 복수의 환자에 대한 진료정보를 입력으로, 상기 복수의 환자에 대해 발생될 수 있는 적어도 하나의 의료 이벤트를 정답으로 하여 딥러닝 학습 모델을 학습시킬 수 있다(단계 S1).Referring to FIG. 6 , first, in order to learn the deep learning learning model, the processor 120 receives medical information about a plurality of patients as an input, and at least one medical event that may be generated for the plurality of patients as a correct answer. A deep learning learning model may be trained (step S1).
이후, 딥러닝 학습모델의 정확성을 평가하기 위해, 프로세서(120)는 딥러닝 학습 모델에 복수의 환자에 대한 진료정보 중 일부를 입력으로 넣고, 원하는 의료 이벤트가 출력되는지를 확인할 수 있다(단계 S2).Thereafter, in order to evaluate the accuracy of the deep learning learning model, the processor 120 may input some of the medical information for a plurality of patients to the deep learning learning model as input, and check whether a desired medical event is output (step S2). ).
이때, 프로세서(120)는 딥러닝 학습모델에서 출력되는 의료 이벤트의 오차값이 기 설정된 값 이하인지 확인(단계 S3)할 수 있으며, 딥러닝 학습모델에서 출력되는 의료 이벤트의 오차값이 기 설정된 값을 초과할 경우, 딥러닝 학습모델은 다시 단계 S2에서 학습될 수 있다.At this time, the processor 120 may check whether the error value of the medical event output from the deep learning learning model is equal to or less than a preset value (step S3), and the error value of the medical event output from the deep learning learning model is a preset value. is exceeded, the deep learning learning model may be trained again in step S2.
한편, 프로세서(120)는 딥러닝 학습모델에서 출력되는 의료 이벤트의 오차값이 기 설정된 값 이하일 경우에는, 딥러닝 학습모델이 메모리(110) 또는 외부의 다른 장치에 저장될 수 있다(단계 S4).On the other hand, the processor 120, when the error value of the medical event output from the deep learning learning model is less than or equal to a preset value, the deep learning learning model may be stored in the memory 110 or other external device (step S4) .
이후, 프로세서(120)는 새로운 데이터(예를 들어, 환자의 진료정보)가 유입되었는지 확인(단계 S5)할 수 있으며, 이때 새로운 데이터(예를 들어, 환자의 진료정보)가 유입될 경우에는 단계 S1으로 가서, 딥러닝 학습모델을 학습시키기 위한 입력 데이터로 새로운 데이터를 입력할 수 있다. 하지만, 새로운 데이터가 유입되지 않을 경우에는 종료될 수 있다.Thereafter, the processor 120 may check whether new data (eg, patient's medical information) is introduced (step S5), and in this case, when new data (eg, patient's medical information) is introduced, step Go to S1, you can input new data as input data for training the deep learning learning model. However, it may be terminated if no new data is introduced.
한편, 도 7에서 도시된 바와 같이, 딥러닝 학습모델의 입력데이터로서, 복수의 환자에 대한 진료정보가 입력될 수 있으며, 이때 복수의 환자에 대한 진료정보는 환자가 현재 진료정보를 획득한 병원(기관 A) 이외의 타 병원에서의 진료정보(기관 B)를 포함할 수 있다.Meanwhile, as shown in FIG. 7 , as input data of the deep learning learning model, treatment information for a plurality of patients may be input, and in this case, the treatment information for the plurality of patients is the hospital in which the patient has acquired the current treatment information. It may include medical information from other hospitals other than (Institution A) (Institution B).
다시 도 1을 참조하여, 프로세서(120)는 환자의 진료정보를 딥러닝 학습모델에 입력하여 환자에 대한 적어도 하나의 의료 이벤트를 예측할 수 있다.Referring back to FIG. 1 , the processor 120 may predict at least one medical event for the patient by inputting the patient's medical treatment information into the deep learning learning model.
여기서, 의료 이벤트는 환자의 퇴원 후 소정기간 이내 사망률, 환자의 퇴원 후 소정기간(예를 들어, 30일일 수 있음) 이내 재입원률, 환자가 재입원한 시점에서부터 퇴원시점까지의 기간 및 환자의 퇴원 후 소정기간(예를 들어, 30일일 수 있음) 이내 심장병의 발병률 중 적어도 하나를 포함할 수 있다.Here, the medical event includes the mortality rate within a predetermined period after the patient's discharge, the readmission rate within a predetermined period (for example, may be 30 days) after the patient's discharge, the period from the time the patient is re-admitted to the time of discharge, and the discharge of the patient. It may include at least one of the incidence rates of heart disease within a predetermined period (eg, may be 30 days) after.
보다 상세히, 환자의 퇴원 후 소정기간 이내 사망률은 원내 사망(in-hospital death) 및 암에 의한 사망을 포함할 수 있다.In more detail, the mortality rate within a predetermined period after discharge of a patient may include in-hospital death and death due to cancer.
또한, 환자의 퇴원 후 소정기간 이내 사망률은 하기 표 1에 도시된 바와 같이, 병원 내의 전체 환자수(전체 입원 건수)와 병원 내의 전체 환자 중 퇴원 시점에서 30일 이내 사망 환자수(퇴원 시점에서 30일 이내 사망 환자수)를 이용하여 계산될 수 있다.In addition, as shown in Table 1 below, the mortality rate within a predetermined period after the patient's discharge was the total number of patients in the hospital (total number of hospitalizations) and the number of patients who died within 30 days from the time of discharge (30 at the time of discharge) among all patients in the hospital. number of deaths within one day).
전체 입원 건수Total number of hospitalizations 1,987,033 건1,987,033 cases
퇴원 시점에서 30일 이내 사망 환자수Number of patients who died within 30 days of discharge 1754건(4.2%)1754 cases (4.2%)
AUROCAUROC 0.92390.9239
이때, 상기 표 1에서 AUROC는 딥러닝 학습모델을 통해 예측된 환자의 퇴원 후 소정기간 이내 사망률의 정확도를 나타낸 값이다. 예를 들어, 딥러닝 학습모델을 통해 예측된 환자의 퇴원 후 소정기간 이내 사망률의 정확도는 AUROC의 값에 100을 곱하여, 92.39%라고 할 수 있다.In this case, in Table 1, AUROC is a value indicating the accuracy of mortality within a predetermined period after discharge of a patient predicted through a deep learning learning model. For example, the accuracy of mortality within a predetermined period after discharge of a patient predicted through a deep learning learning model is 92.39% by multiplying the value of AUROC by 100.
환자의 퇴원 후 소정기간(예를 들어, 30일일 수 있음) 이내 재입원률은 입원 전과 입원 중의 환자들의 진료 정보를 고려하여 환자의 퇴원 시점에서 30일 이내 재입원에 대한 정보이다. 이때, 재입원은 병원에 한번 이상 입원한 기록이 있는 환자에 한하여 퇴원 시점에서 다시 입원한 시간이 30일 이내인 경우를 의미한다.The readmission rate within a predetermined period (for example, may be 30 days) after a patient's discharge is information about readmission within 30 days from the patient's discharge in consideration of the patient's medical information before and during hospitalization. In this case, readmission refers to a case where the time of re-admission from the time of discharge is within 30 days only for patients who have been admitted to the hospital at least once.
또한, 환자의 퇴원 후 소정기간(예를 들어, 30일일 수 있음) 이내 재입원된 환자 중 장기 입원된 환자율은 하기 표 2에 도시된 바와 같이, 병원 내의 전체 환자 중 30일 이내 재입원한 환자의 수(30일 이내 재입원 건수)와 병원 내의 전체 환자 중 재입원 시점부터 퇴원 시점까지의 기간이 7일을 초과한 환자의 수(재입원 중 장기 입원 건수 (7일 초과))를 이용하여 계산할 수 있다.In addition, as shown in Table 2 below, the rate of long-term hospitalized patients among patients re-hospitalized within a predetermined period (for example, may be 30 days) after the patient's discharge was within 30 days among all patients in the hospital. Using the number of patients (the number of readmissions within 30 days) and the number of patients whose period from readmission to discharge exceeds 7 days (the number of long-term hospitalizations during readmission (more than 7 days)) among all patients in the hospital can be calculated by
30일 이내 재입원 건수Number of readmissions within 30 days 14,223건14,223 cases
재입원 중 장기 입원 건수(7일초과)Number of long-term hospitalizations during readmission (more than 7 days) 5,300건(37.26%)5,300 cases (37.26%)
AUROCAUROC 0.81260.8126
이때, 상기 표 2에서의 AUROC는 딥러닝 학습모델을 통해 예측된 환자의 퇴원 후 소정기간(예를 들어, 30일일 수 있음) 이내 재입원된 환자 중 장기 입원된 환자율의 정확도를 나타낸 값이다. 예를 들어, 딥러닝 학습모델을 통해 환자의 퇴원 후 소정기간(예를 들어, 30일일 수 있음) 이내 재입원된 환자 중 장기 입원된 환자율의 정확도는 AUROC의 값에 100을 곱하여, 81.26%라고 할 수 있다.In this case, AUROC in Table 2 is a value indicating the accuracy of the long-term hospitalization rate among patients re-hospitalized within a predetermined period (for example, it may be 30 days) after discharge of the patient predicted through the deep learning learning model. . For example, the accuracy of the long-term hospitalization rate among patients re-hospitalized within a predetermined period (for example, 30 days) after patient discharge through a deep learning learning model is 81.26% by multiplying the value of AUROC by 100 It can be said that
아래 표 3에서는, 환자가 퇴원한 후 30일 이내에 재입원한 환자에 대한 데이터를 나타내며, 그 중 환자가 재입원한 시점에서부터 퇴원시점까지의 기간이 소정 기간 이하인지 여부를 나타내는 표이다. 여기서 소정 기간은 10일로 예시적으로 표현되어 있다. Table 3 below shows data on patients who were readmitted within 30 days after the patient was discharged, and among them, it is a table indicating whether the period from the time the patient was readmitted to the time of discharge was less than or equal to a predetermined period. Here, the predetermined period is exemplarily expressed as 10 days.
또한, 환자가 재입원한 시점에서부터 퇴원시점까지의 기간은 하기 표 3에 도시된 바와 같이, 병원 내의 전체 환자 중 30일 이내 재입원한 환자의 수(30일 이내 재입원 건수)와 병원 내의 전체 환자 중 30일 이내 재입원한 환자 중 10일 미만의 기간으로 입원한 환자의 수(그 중 10일 미만 입원한 건수)를 이용하여 계산할 수 있다.In addition, as shown in Table 3 below, the period from the time the patient was re-admitted to the time of discharge is the number of patients who were readmitted within 30 days (the number of readmissions within 30 days) and the total number of patients in the hospital, as shown in Table 3 below. It can be calculated using the number of patients who were readmitted within 30 days and who were hospitalized for less than 10 days (the number of hospitalizations for less than 10 days among them).
30일 이내 재입원 건수Number of readmissions within 30 days 14,223건14,223 cases
그 중 10일 미만 입원한 건수Number of hospitalizations of less than 10 days 10,285건(72.31%)10,285 cases (72.31%)
R^2R^2 0.23920.2392
Mean Absolute ErrorMean Absolute Error 1.76321.7632
이때, 상기 표 3에서 R^2은 입원일수 혹은 입원기간을 얼마나 정확하게 예측했는지를 확인할 수 있는 지수이다. 또한, Mean Absolute Error는 예측된 환자가 재입원한 시점에서부터 퇴원시점까지의 기간의 오차율을 의미할 수 있다.In this case, in Table 3, R^2 is an index that can confirm how accurately the number of hospitalization days or hospitalization period was predicted. In addition, the mean absolute error may mean the error rate of the period from the time when the predicted patient is re-hospitalized to the time of discharge.
환자의 퇴원 후 소정기간(예를 들어, 30일일 수 있음) 이내 심장병의 발병률은 병원 내의 전체 환자 중 퇴원한 후에 심장병이 30일 이내에 발생하는지를 예측한 이벤트이다. 여기서, 심장병은 사망, 재개통(revascularization), 심근경색, 뇌졸중 중 적어도 하나를 포함하며, 30일 이내에 가장 먼저 발생한 사건이 고려된다.The incidence rate of heart disease within a predetermined period (eg, may be 30 days) after a patient's discharge is an event that predicts whether a heart disease will occur within 30 days after discharge among all patients in the hospital. Here, the heart disease includes at least one of death, revascularization, myocardial infarction, and stroke, and the earliest event within 30 days is considered.
또한, 환자의 퇴원 후 소정기간(예를 들어, 30일일 수 있음) 이내 심장병의 발병률은 하기 표 4에 도시된 바와 같이, 병원 내의 전체 환자 중 퇴원 후 다시 병원을 방문한 환자의 수(전체 방문 건수)와 병원 내의 전체 환자 중 심장병(예를 들어, 사망, 재개통(revascularization), 심근경색, 뇌졸중)이 발생한 환자의 수(그 중 주요 심장 사건이 발생한 건수)를 이용하여 계산할 수 있다. In addition, as shown in Table 4 below, the incidence rate of heart disease within a predetermined period (for example, may be 30 days) after the patient's discharge is the number of patients who visited the hospital again after being discharged from the hospital (the total number of visits) ) and the total number of patients in the hospital with heart disease (eg, death, revascularization, myocardial infarction, stroke) (the number of major cardiac events).
전체 방문 건수total number of visits 922,580922,580
그 중 주요 심장 사건이 발생한 건수number of major cardiac events 10,02410,024
AUROCAUROC 0.79870.7987
한편, 의료 이벤트를 예측하는데 사용되는 딥러닝 학습모델은 부스팅된 시간-인식(Time-aware) 임베딩(또는 피처(feature))를 갖는 피드 포워드 모델(FFNN: Feed-forward Neural Network)과 그래디언트 부스팅 트리 모델(GBM: Gradient Boosting Machine) 중 적어도 하나의 모델일 수 있다.On the other hand, the deep learning learning model used to predict medical events is a feed-forward model (FFNN: Feed-forward Neural Network) with boosted time-aware embeddings (or features) and a gradient boosting tree. It may be at least one model among models (GBM: Gradient Boosting Machine).
프로세서(120)는 예측된 환자의 의료 이벤트, 입출력부(101)에서 입력 받은 환자에 대한 진료정보 및 병원(환자가 진료받은 병원일 수 있음)의 상태정보를 고려하여 병원의 의사결정을 수행할 수 있다.The processor 120 performs hospital decision-making in consideration of the predicted patient's medical event, medical information about the patient input from the input/output unit 101, and state information of the hospital (which may be the hospital where the patient has been treated). can
여기서, 병원의 상태정보는 병원의 진료실 개수, 병원의 의료진 별 사용한 진료실의 개수, 의료진 별 외래 지원 인력정보, 병원의 동일 진료과 내 타 의료진 요일별 세션 개설 상태정보, 병원의 병실 침대정보, 병원의 재원 환자 상태정보, 진료과 별 상태정보, 수술실의 종류와 개수, 수술실별 스케줄 정보 및 환자의 수술별 소요예정시간 정보 중 적어도 하나를 포함할 수 있으나, 이에 한정되는 것은 아니다.Here, the hospital status information includes the number of clinics in the hospital, the number of treatment rooms used by each medical staff in the hospital, outpatient support personnel information by medical staff, session opening status information by day of the week for other medical staff in the same department of the hospital, bed information in the hospital, and hospital resources It may include at least one of patient status information, status information for each department, type and number of operating rooms, schedule information for each operating room, and estimated time required for each operation of the patient, but is not limited thereto.
병원의 의사결정은 환자의 진료 스케줄 결정, 환자의 수술실 스케줄 결정 및 환자의 병상배정 결정 중 적어도 하나를 포함할 수 있다.The decision-making of the hospital may include at least one of a patient's treatment schedule determination, a patient's operating room schedule determination, and a patient's bed assignment determination.
일 실시예로서, 프로세서(120)는 환자에 대한 진료정보 및 딥러닝 학습 모델을 이용하여 예측된 환자의 의료 이벤트를 기초로 환자의 진료순위를 결정할 수 있다.As an embodiment, the processor 120 may determine the patient's treatment priority based on the patient's medical event predicted using the patient's medical treatment information and a deep learning learning model.
예를 들어, 프로세서(120)는 딥러닝 학습 모델을 이용하여 예측된 환자의 의료 이벤트, 환자에 대한 진료정보, 병원의 상태정보, 환자의 진료순위 및 병원의 의료진 정보를 고려하여 병원의 의사결정을 수행할 수 있다.For example, the processor 120 makes a hospital decision in consideration of the patient's medical event predicted using the deep learning learning model, the patient's medical information, the hospital's status information, the patient's medical priority, and the hospital's medical staff information. can be performed.
여기서, 병원의 의료진 정보는 의료진 별 진료 스케줄정보, 환자 유형별 정보, 환자의 진료에 의해 발생되는 시간정보, 의료진 별 수술, 시술 및 연구 중 적어도 하나의 스케줄 정보, 의료진 별 실적정보 및 병원의 의료진 보직 여부 정보 중 적어도 하나를 포함할 수 있으나, 이에 한정되는 것은 아니다.Here, the medical staff information of the hospital includes treatment schedule information by medical staff, information by patient type, time information generated by patient treatment, schedule information of at least one of surgery, procedure and research by medical staff, performance information by medical staff, and medical staff position in the hospital It may include at least one of whether or not information, but is not limited thereto.
또 다른 실시예로서, 프로세서(120)는 환자에 대한 진료정보 및 딥러닝 학습 모델을 이용하여 예측된 환자의 의료 이벤트를 기초로 환자의 병상배정 순위를 결정할 수 있다.As another embodiment, the processor 120 may determine the patient's bed allocation ranking based on the patient's medical event predicted using the patient's medical treatment information and a deep learning learning model.
예를 들어, 프로세서(120)는 딥러닝 학습 모델을 이용하여 예측된 환자의 의료 이벤트, 환자에 대한 진료정보, 병원의 상태정보 및 환자의 병상(또는 병실)배정 순위를 고려하여 병원의 의사결정을 수행할 수 있다.For example, the processor 120 makes a hospital decision in consideration of the patient's medical event predicted using the deep learning learning model, the medical information about the patient, the state information of the hospital, and the patient's bed (or ward) assignment order can be performed.
이하, 일 실시예에 따른 병원의 의사결정 장치(100)를 이용하여 병원의 의사결정을 수행하는 예(또는 시나리오)를 설명하도록 한다.Hereinafter, an example (or scenario) of performing hospital decision-making using the hospital decision-making apparatus 100 according to an embodiment will be described.
도 8은 일 실시예에 따른 병원의 의사결정을 수행하는 예를 설명하기 위한 흐름도이다. 8 is a flowchart illustrating an example of performing decision-making in a hospital according to an exemplary embodiment.
도 8을 참조하면, 일 실시예에 따른 병원의 의사결정 장치(100)는 먼저 입출력부(101)를 통해 환자의 진료정보를 입력 받을 수 있다(단계 S11). Referring to FIG. 8 , the hospital decision-making apparatus 100 according to an embodiment may first receive the patient's medical treatment information through the input/output unit 101 (step S11).
이때, 일 실시예에 따른 병원의 의사결정 장치(100)는 환자의 진료정보에 환자가 진료정보를 획득한 병원 외 타 병원을 포함한 과거의 환자의 진료정보(예를 들어, 환자의 입원내역 기록)가 존재하지 여부를 확인할 수 있다(단계 S12).In this case, the hospital decision-making apparatus 100 according to an embodiment includes the patient's medical information in the past, including hospitals other than the hospital where the patient obtained the medical information (for example, the patient's hospitalization history record) ) can be checked whether or not there is (step S12).
만약 환자의 환자의 진료정보에 환자가 진료정보를 획득한 병원 외 타 병원을 포함한 과거의 환자의 진료정보(예를 들어, 환자의 입원내역 기록)가 존재하지 않으면, 환자의 진료 순위를 유지할 수 있다. If the patient's medical information does not contain the patient's past medical information (for example, the patient's hospitalization history record) including hospitals other than the hospital from which the patient obtained the medical information, the patient's medical priority cannot be maintained. have.
이때, 환자의 초기 진료순위는 병원 내 전체 환자의 진료순위 중 제일 마지막 순위로 설정되어 있을 수 있으나, 이에 한정되는 것은 아니다.In this case, the initial treatment order of the patient may be set as the last priority among the treatment orders of all patients in the hospital, but is not limited thereto.
이후, 병원의 의사결정 장치(100)는 환자의 진료정보에 환자가 진료정보를 획득한 병원 외 타 병원을 포함한 과거의 환자의 진료정보(예를 들어, 환자의 입원내역 기록)가 존재하면, 환자의 진료정보에 대하여 딥러닝 학습모델을 통해 의료 이벤트 중 환자의 퇴원 후 소정기간 이내 사망률을 예측하고, 예측된 환자의 퇴원 후 소정기간 이내 사망률이 기 설정된 확률을 초과하는지 확인할 수 있다(단계 S13).Thereafter, the hospital decision-making device 100 includes past patient medical information (eg, patient's hospitalization history record) including hospitals other than the hospital where the patient obtained medical information in the patient's medical information. It is possible to predict the mortality rate within a predetermined period after a patient's discharge during a medical event through a deep learning learning model on the patient's medical information, and check whether the predicted mortality rate within a predetermined period after discharge exceeds a preset probability (step S13) ).
이 경우, 예측된 환자의 퇴원 후 소정기간 이내 사망률이 기 설정된 확률 이하이면 환자의 진료 순위는 그대로 유지될 수 있다.In this case, if the predicted mortality rate within a predetermined period after discharge is less than or equal to a preset probability, the patient's treatment order may be maintained as it is.
이때, 병원의 의사결정 장치(100)는 예측된 환자의 퇴원 후 소정기간 이내 사망률이 기 설정된 확률을 초과할 경우, 환자의 진료 순위는 상승할 수 있다(단계 S14).In this case, in the hospital decision-making apparatus 100, when the predicted mortality rate within a predetermined period after discharge exceeds a preset probability, the patient's medical treatment order may be increased (step S14).
이후, 환자의 진료 순위에 따라 일 실시예에 따른 병원의 의사결정 장치(100)는 예측된 환자의 의료 이벤트, 환자에 대한 진료정보, 병원의 상태정보, 환자의 진료순위 및 병원의 의료진 정보를 고려하여 병원의 의사결정 중 환자의 진료 스케줄을 결정할 수 있다(단계 S15). Thereafter, according to the patient's treatment order, the hospital decision-making apparatus 100 according to an embodiment provides the predicted patient's medical event, the patient's medical information, the hospital's state information, the patient's medical priority, and the hospital's medical staff information. Taking this into account, it is possible to determine the patient's treatment schedule during the decision-making of the hospital (step S15).
이때, 단계 S15에서 환자의 진료 스케줄을 결정하는데 있어 사용되는 정보 중 환자에 대한 진료정보는 환자의 퇴원 후 소정 기간 이내 사망률 및 병원의 방문 기준 시점 이전의 병원 방문 횟수, 진료 대기시간 및 진료 소요시간, 진단 정보, 투약력 정보, 수술/시술/처치 정보 및 신체계측 정보 중 적어도 하나의 정보를 포함한 병원 방문정보 중 적어도 하나의 정보를 포함할 수 있다. At this time, among the information used to determine the patient's treatment schedule in step S15, the treatment information about the patient includes the mortality rate within a predetermined period after the patient's discharge, the number of hospital visits before the reference point of the hospital visit, the waiting time for treatment, and the time required for treatment. , diagnostic information, medication history information, surgery/procedure/treatment information, and may include at least one information of hospital visit information including at least one of body measurement information.
또한, 의료진측 정보는 진료 스케줄(예를 들어, 주별 개설 횟수, 요일, 오전오후, 총 진료시간, 타임당 환자 정원, 외래 회송률(휴진율, 추가진료율)정보일 수 있음), 환자 유형별 정보(예를 들어, 환자 유형비율, 정원, 재진료율, 예약율, 예약 대비 실진료율, 미내원율, 중증환자의 비율 정보일 수 있음), 환자 별 평균 대기 시간, 환자의 진료 소요시간, 환자 종료 지연시간정보, 의료진 수술/시술/연구의 스케쥴, 수술/시술/연구의 실적 건수, 외래 후 수술 연계비율 및 보직여부 정보 중 적어도 하나의 정보를 포함할 수 있다.In addition, the medical staff information may be the treatment schedule (for example, the number of openings per week, days of the week, morning and afternoon, total treatment hours, number of patients per time, outpatient return rate (discontinuance rate, additional treatment rate) information), by patient type Information (for example, patient type ratio, capacity, re-treatment rate, reservation rate, actual treatment rate compared to reservation, non-visiting rate, ratio of critically ill patients may be information), average waiting time for each patient, patient's treatment time, It may include information about at least one of patient termination delay time information, medical staff surgery/procedure/research schedule, number of surgical/procedure/research results, post-outpatient surgery linkage rate, and position status information.
또한, 병원 상태정보는 진료과의 진료실 개수, 의료진별/진료과별 진료실 사용 개수, 의료진별 외래 지원 인력 및 동일 진료과 내 타 의료진 요일별 세션 개설 현황 정보 중 적어도 하나의 정보를 포함할 수 있다.In addition, the hospital status information may include at least one of the number of clinics in each department, the number of offices used by medical staff/department, outpatient support personnel for each medical team, and session opening status information by day of the week for other medical staff in the same department.
한편, 단계 S15 이후, 환자는 결정된 환자의 진료스케줄에 따라 진료를 받을 수 있다.Meanwhile, after step S15, the patient may receive treatment according to the determined patient's treatment schedule.
도 9는 또 다른 실시예에 따른 병원의 의사결정을 수행하는 예를 설명하기 위한 흐름도이다.9 is a flowchart illustrating an example of performing decision-making in a hospital according to another embodiment.
도 9를 참조하면, 먼저, 일 실시예에 따른 병원의 의사결정 장치(100)는 환자에 대한 진료정보 중 입원 예약 환자 리스트에 대한 정보를 입력 받을 수 있다(단계 S21).Referring to FIG. 9 , first, according to an embodiment The hospital decision-making apparatus 100 may receive information about a list of patients reserved for hospitalization among medical information about the patient (step S21).
이후, 일 실시예에 따른 병원의 의사결정 장치(100)는 환자의 진료정보(예약 환자 리스트 정보)에 대하여 딥러닝 학습모델을 통해 의료 이벤트 중 환자의 퇴원 후 소정기간 이내 심장병의 발병률을 예측하고, 예측된 환자의 퇴원 후 소정기간 이내 심장병의 발병률이 기 설정된 확률을 초과하는지 확인(단계 S22)할 수 있으며, 예측된 환자의 퇴원 후 소정기간 이내 심장병의 발병률이 기 설정된 확률 이하일 경우, 환자의 병상(또는 병실)배정 순위는 그대로 유지될 수 있다.Thereafter, the hospital decision-making apparatus 100 according to an embodiment predicts the incidence rate of heart disease within a predetermined period after the patient's discharge during a medical event through a deep learning learning model with respect to the patient's medical information (reserved patient list information), , it is possible to check whether the predicted incidence of heart disease within a predetermined period after discharge exceeds a preset probability (step S22), and when the predicted incidence of heart disease within a predetermined period after discharge of the patient is less than or equal to a preset probability, the patient's The hospital bed (or ward) allocation order may be maintained as it is.
이때, 환자의 초기 병상(또는 병실)배정 순위는 병원 내 전체 환자의 병상(또는 병실)배정 순위 중 제일 마지막 순위로 설정되어 있을 수 있으나, 이에 한정되는 것은 아니다.In this case, the initial bed (or ward) assignment order of the patient may be set to the last rank among the bed (or room) assignment orders of all patients in the hospital, but is not limited thereto.
한편, 병원의 의사결정 장치(100)는 예측된 환자의 퇴원 후 소정기간 이내 심장병의 발병률이 기 설정된 확률을 초과할 경우, 환자의 병상(또는 병실)배정 순위는 상승할 수 있다(단계 S23). On the other hand, in the hospital decision-making apparatus 100, when the predicted incidence rate of heart disease within a predetermined period after discharge exceeds a preset probability, the patient's bed (or ward) assignment order may rise (step S23). .
이후, 일 실시예에 따른 병원의 의사결정 장치(100)는 환자의 병상(또는 병실)배정 순위에 따라 예측된 환자의 의료 이벤트, 환자에 대한 진료정보, 병원의 상태정보 및 환자의 병상(또는 병실)배정 순위를 고려하여 병원의 의사결정 중 환자의 병상배정을 결정할 수 있다(단계 S24).Thereafter, the hospital decision-making apparatus 100 according to an embodiment may provide a patient's medical event predicted according to the patient's bed (or room) assignment order, medical treatment information for the patient, hospital status information, and the patient's bed (or The patient's bed assignment may be determined during decision-making in the hospital by considering the ward) assignment order (step S24).
이때, 단계 S24에서 환자의 병상배정을 결정하는데 있어 사용되는 정보 중 환자에 대한 진료정보는 환자의 등록번호(또는 내원번호), 성명, 성별, 나이, 환자의 유형(신/초/재진) 등을 포함한 기본 정보, 진단명, 진료과, 담당의 등을 포함한 진단 정보, 중증등급, 주의사항, 특기사항, 격리종류/접수일/필요기간 등을 포함한 환자의 상태정보, 시술/수술명, 수술예정일자, 집도의, 마취종류 등을 포함한 환자의 수술 정보, 입원기간내의 약, 재료의 종류/처방/처치 등을 포함한 환자의 처치 정보 및 각종 진단검사 및 영상검사의 결과 정보 중 적어도 하나의 정보를 포함할 수 있다.At this time, among the information used to determine the patient's bed assignment in step S24, the medical information about the patient includes the patient's registration number (or visit number), name, gender, age, patient type (new/secondary/revisit), etc. Basic information including, diagnosis information, including name of diagnosis, department, doctor in charge, severity level, precautions, special notes, patient status information including type of isolation/receipt date/required period, operation/surgery name, scheduled surgery date, At least one of information about the patient's surgery including the surgeon and the type of anesthesia, the medicine within the hospital stay, the patient's treatment information including the type/prescription/treatment of materials, and the result information of various diagnostic tests and imaging tests. can
또한, 병원 상태 정보는 병동, 병실, 병상의 이름/번호/개수 등을 포함한 전체 침대(total bed)의 정보, 현재 모든 재원 환자가 사용하고 있는 침대(bed)의 병동/병실/병상 별 개수, 재원 환자 상태(입원기간, 수술/시술/처치 여부 등), 진료과 별 병동/병실/병상 현황, 수술실 종류와 개수, 수술실 별 스케줄 현황 (수술명, 수술 시간표) 및 수술 별 수술 소요예정시간 중 적어도 하나의 정보를 포함할 수 있다.In addition, hospital status information includes total bed information including ward, ward, name/number/number of beds, etc., the number of beds used by all hospitalized patients per ward/ward/bed, At least among inpatient patient status (hospitalization period, surgery/procedure/treatment status, etc.), ward/ward/bed status by department, type and number of operating rooms, schedule status by operating room (surgery name, surgery timetable), and scheduled operation time for each surgery It may contain one piece of information.
더 나아가, 단계 S24에서 환자의 병상배정을 결정하는데 있어서, 입원 시점에 입원 기간 예측 값이 31일 이상인 총 환자의 수(입원 기간이 장기간인 총 환자의 수), 입원 시점에 입원 기간 예측 값이 30일 보다 작은 소정 값 N일 이내로 입원기간이(예를 들어, N은 3일) 예측되는 총 환자의 수(입원 기간이 단기간인 총 환자의 수) 및 30일 이내 재 입원 가능성이 있는 총 환자의 수의 정보 중 적어도 하나의 정보가 고려될 수 있으나, 이에 한정되는 것은 아니다.Further, in determining the bed assignment of the patient in step S24, the total number of patients with a predicted hospitalization period of 31 days or more at the time of admission (the total number of patients with a long hospitalization period), the predicted value of the hospitalization period at the time of admission Predetermined value less than 30 days Total number of patients expected to be hospitalized within N days (eg, N is 3 days) (total number of short-stay patients) and total patients likely to be readmitted within 30 days At least one piece of information among the number of pieces of information may be considered, but is not limited thereto.
한편, 단계 S24 이후 환자의 병상 배정이 성공 또는 실패할 수 있고, 환자의 병상 배정이 성공할 경우, 환자는 결정된 병상 배정에 따라 입원절차를 수행할 수 있다.Meanwhile, after step S24, the patient's bed assignment may succeed or fail, and when the patient's bed assignment succeeds, the patient may perform a hospitalization procedure according to the determined bed assignment.
도 10은 일 실시예에 따른 병원의 의사결정 방법에 대한 예시적인 흐름도이다. 도 10에 도시된 병원의 의사결정 방법은 도 1에 도시된 병원의 의사결정 장치(100)에 의해 수행 가능하다. 아울러, 도 10에 도시된 병원의 의사결정 방법은 예시적인 것에 불과하다.10 is an exemplary flowchart of a decision-making method in a hospital according to an embodiment. The hospital decision-making method shown in FIG. 10 can be performed by the hospital decision-making device 100 shown in FIG. 1 . In addition, the hospital decision-making method shown in FIG. 10 is merely exemplary.
먼저, 일 실시예에 따른 병원의 의사결정 장치(100)는 복수의 환자에 대한 진료정보를 기 정해진 구조 형식의 데이터로 변환시킬 수 있다(단계 S100).First, the hospital decision-making apparatus 100 according to an embodiment may convert medical treatment information for a plurality of patients into data having a predetermined structure (step S100).
이후, 병원의 의사결정 장치(100)는 기 정해진 구조 형식의 데이터를 입력으로, 복수의 환자에 대해 발생될 수 있는 적어도 하나의 의료 이벤트를 정답으로 하여 딥러닝 학습모델을 학습시킬 수 있다(단계 S200).Thereafter, the hospital decision-making apparatus 100 may train the deep learning learning model by using data of a predetermined structural format as an input, and at least one medical event that may be generated for a plurality of patients as a correct answer (step S200).
이후, 병원의 의사결정 장치(100)는 환자에 대한 진료정보를 입력 받아, 딥러닝 학습 모델을 통해 환자에 대한 적어도 하나의 의료 이벤트를 예측할 수 있다(단계 S300).Thereafter, the hospital decision-making apparatus 100 may receive medical treatment information for the patient and predict at least one medical event for the patient through the deep learning learning model (step S300 ).
이후, 병원의 의사결정 장치(100)는 예측된 환자에 대한 적어도 하나의 의료 이벤트, 환자에 대한 진료정보 및 병원의 상태정보를 고려하여 병원의 의사결정을 수행할 수 있다(단계 S400).Thereafter, the hospital decision-making apparatus 100 may perform the hospital decision-making in consideration of the predicted at least one medical event for the patient, medical treatment information for the patient, and state information of the hospital (step S400 ).
이상에서 살펴본 바와 같이, 일 실시예에 따르면 병원의 의사결정 장치는 딥러닝 학습 모델을 이용하여 의료 이벤트(예를 들어, 환자의 퇴원 후 소정기간 이내 사망률, 환자의 퇴원 후 소정기간 이내 재입원률, 환자가 재입원한 시점에서부터 퇴원시점까지의 기간 및 환자의 퇴원 후 소정기간 이내 심장병의 발병률)를 예측하고, 예측된 의료 이벤트, 환자에 대한 진료정보 및 병원의 상태정보를 종합적으로 고려하여 병원 내의 시스템을 관리할 수 있기 때문에 환자 맞춤형 병원 프로세스를 제공할 수 있다.As described above, according to an embodiment, the hospital decision-making device uses a deep learning learning model to perform medical events (eg, mortality within a predetermined period after a patient's discharge, a readmission rate within a predetermined period after a patient's discharge, (the period from the patient's re-hospitalization to the time of discharge and the incidence rate of heart disease within a predetermined period after the patient's discharge) Because the system can be managed, it is possible to provide patient-specific hospital processes.
본 발명에 첨부된 블록도의 각 블록과 흐름도의 각 단계의 조합들은 컴퓨터 프로그램 인스트럭션들에 의해 수행될 수도 있다. 이들 컴퓨터 프로그램 인스트럭션들은 범용 컴퓨터, 특수용 컴퓨터 또는 기타 프로그램 가능한 데이터 프로세싱 장비의 인코딩 프로세서에 탑재될 수 있으므로, 컴퓨터 또는 기타 프로그램 가능한 데이터 프로세싱 장비의 인코딩 프로세서를 통해 수행되는 그 인스트럭션들이 블록도의 각 블록 또는 흐름도의 각 단계에서 설명된 기능들을 수행하는 수단을 생성하게 된다. 이들 컴퓨터 프로그램 인스트럭션들은 특정 방법으로 기능을 구현하기 위해 컴퓨터 또는 기타 프로그램 가능한 데이터 프로세싱 장비를 지향할 수 있는 컴퓨터 이용 가능 또는 컴퓨터 판독 가능 메모리에 저장되는 것도 가능하므로, 그 컴퓨터 이용가능 또는 컴퓨터 판독 가능 메모리에 저장된 인스트럭션들은 블록도의 각 블록 또는 흐름도 각 단계에서 설명된 기능을 수행하는 인스트럭션 수단을 내포하는 제조 품목을 생산하는 것도 가능하다. 컴퓨터 프로그램 인스트럭션들은 컴퓨터 또는 기타 프로그램 가능한 데이터 프로세싱 장비 상에 탑재되는 것도 가능하므로, 컴퓨터 또는 기타 프로그램 가능한 데이터 프로세싱 장비 상에서 일련의 동작 단계들이 수행되어 컴퓨터로 실행되는 프로세스를 생성해서 컴퓨터 또는 기타 프로그램 가능한 데이터 프로세싱 장비를 수행하는 인스트럭션들은 블록도의 각 블록 및 흐름도의 각 단계에서 설명된 기능들을 실행하기 위한 단계들을 제공하는 것도 가능하다.Combinations of each block in the block diagram attached to the present invention and each step in the flowchart may be performed by computer program instructions. These computer program instructions may be embodied in the encoding processor of a general purpose computer, special purpose computer, or other programmable data processing equipment, such that the instructions executed by the encoding processor of the computer or other programmable data processing equipment may correspond to each block of the block diagram or Each step of the flowchart creates a means for performing the functions described. These computer program instructions may also be stored in a computer-usable or computer-readable memory which may direct a computer or other programmable data processing equipment to implement a function in a particular way, and thus the computer-usable or computer-readable memory. The instructions stored in the block diagram may also produce an item of manufacture containing instruction means for performing a function described in each block of the block diagram or each step of the flowchart. The computer program instructions may also be mounted on a computer or other programmable data processing equipment, such that a series of operational steps are performed on the computer or other programmable data processing equipment to create a computer-executed process to create a computer or other programmable data processing equipment. It is also possible that instructions for performing the processing equipment provide steps for carrying out the functions described in each block of the block diagram and in each step of the flowchart.
또한, 각 블록 또는 각 단계는 특정된 논리적 기능(들)을 실행하기 위한 하나 이상의 실행 가능한 인스트럭션들을 포함하는 모듈, 세그먼트 또는 코드의 일부를 나타낼 수 있다. 또, 몇 가지 대체 실시예들에서는 블록들 또는 단계들에서 언급된 기능들이 순서를 벗어나서 발생하는 것도 가능함을 주목해야 한다. 예컨대, 잇달아 도시되어 있는 두 개의 블록들 또는 단계들은 사실 실질적으로 동시에 수행되는 것도 가능하고 또는 그 블록들 또는 단계들이 때때로 해당하는 기능에 따라 역순으로 수행되는 것도 가능하다.Further, each block or each step may represent a module, segment, or portion of code comprising one or more executable instructions for executing specified logical function(s). It should also be noted that in some alternative embodiments it is also possible for the functions recited in blocks or steps to occur out of order. For example, it is possible that two blocks or steps shown one after another may in fact be performed substantially simultaneously, or that the blocks or steps may sometimes be performed in the reverse order according to the corresponding function.
이상의 설명은 본 발명의 기술 사상을 예시적으로 설명한 것에 불과한 것으로서, 본 발명이 속하는 기술 분야에서 통상의 지식을 가진 자라면 본 발명의 본질적인 품질에서 벗어나지 않는 범위에서 다양한 수정 및 변형이 가능할 것이다. 따라서, 본 발명에 개시된 실시예들은 본 발명의 기술 사상을 한정하기 위한 것이 아니라 설명하기 위한 것이고, 이러한 실시예에 의하여 본 발명의 기술 사상의 범위가 한정되는 것은 아니다. 본 발명의 보호 범위는 아래의 청구범위에 의하여 해석되어야 하며, 그와 균등한 범위 내에 있는 모든 기술사상은 본 발명의 권리범위에 포함되는 것으로 해석되어야 할 것이다.The above description is merely illustrative of the technical spirit of the present invention, and various modifications and variations will be possible without departing from the essential quality of the present invention by those skilled in the art to which the present invention pertains. Accordingly, the embodiments disclosed in the present invention are not intended to limit the technical spirit of the present invention, but to explain, and the scope of the technical spirit of the present invention is not limited by these embodiments. The protection scope of the present invention should be interpreted by the following claims, and all technical ideas within the scope equivalent thereto should be interpreted as being included in the scope of the present invention.

Claims (17)

  1. 환자에 대한 진료정보를 입력 받는 입출력부;an input/output unit for receiving medical information about a patient;
    복수의 환자에 대한 진료정보가 기 정해진 구조 형식의 데이터로 변환되어 저장되어 있는 메모리; 및a memory in which medical information for a plurality of patients is converted into data of a predetermined structure format and stored; and
    상기 메모리와 전기적으로 연결되는 프로세서를 포함하며,a processor electrically connected to the memory;
    상기 프로세서는,The processor is
    상기 기 정해진 구조 형식의 데이터를 입력으로, 상기 복수의 환자에 대해 발생될 수 있는 적어도 하나의 의료 이벤트를 정답으로하여 학습된, 딥러닝 학습모델을 통해 상기 환자에 대한 적어도 하나의 의료 이벤트를 예측하고, 상기 예측된 상기 환자에 대한 적어도 하나의 의료 이벤트, 상기 환자에 대한 진료정보 및 병원의 상태정보를 고려하여 상기 병원의 의사결정을 수행하는Predict at least one medical event for the patient through a deep learning learning model, which is learned by taking as an answer at least one medical event that can occur for the plurality of patients as an input with the data in the predetermined structural format and performing decision-making of the hospital in consideration of the predicted at least one medical event for the patient, medical treatment information for the patient, and state information of the hospital
    병원의 의사결정 장치.Hospital decision-making equipment.
  2. 제 1 항에 있어서,The method of claim 1,
    상기 환자에 대한 진료정보는,Medical information about the patient,
    상기 환자가 상기 병원을 최초 방문한 시점에 따른 진료정보, 상기 환자가 상기 병원을 최초 방문한 시점 이후에 상기 병원을 방문한 시점에 따른 진료정보 및 상기 병원 이외의 타 병원에서 획득한 진료정보 중 적어도 하나를 포함하는At least one of medical treatment information according to the time when the patient first visited the hospital, medical information according to the time when the patient visited the hospital after the initial visit to the hospital, and medical information obtained from other hospitals other than the hospital containing
    병원의 의사결정 장치.Hospital decision-making equipment.
  3. 제 1 항에 있어서,The method of claim 1,
    상기 의료 이벤트는,The medical event is
    상기 환자의 퇴원 후 소정기간 이내 사망률, 상기 환자의 퇴원 후 소정기간 이내 재입원률, 상기 환자가 재입원한 시점에서부터 퇴원시점까지의 기간 및 상기 환자의 퇴원 후 소정기간 이내 심장병의 발병률 중 적어도 하나를 포함하는At least one of the mortality rate within a predetermined period after the patient's discharge, the readmission rate within a predetermined period after the patient's discharge, the period from the patient's readmission to the time of discharge, and the incidence rate of heart disease within the predetermined period after the patient's discharge containing
    병원의 의사결정 장치.Hospital decision-making equipment.
  4. 제 1 항에 있어서,The method of claim 1,
    상기 병원의 의사결정은,The hospital's decision-making
    상기 환자의 진료 스케줄 결정, 상기 환자의 병상배정 결정 및 상기 환자의 수술실 스케줄 결정을 포함하는Determining the patient's treatment schedule, determining the bed assignment of the patient, and determining the operating room schedule of the patient
    병원의 의사결정 장치.Hospital decision-making equipment.
  5. 제 1 항에 있어서,The method of claim 1,
    상기 병원의 상태정보는,The hospital's status information is
    상기 병원의 진료실 개수, 상기 병원의 의료진 별 사용한 진료실의 개수, 상기 의료진 별 외래 지원 인력정보, 상기 병원의 동일 진료과 내 타 의료진 요일별 세션 개설 상태정보, 상기 병원의 병실 침대정보, 상기 병원의 재원 환자 상태정보, 상기 진료과 별 상태정보, 상기 병원의 수술실의 종류와 개수, 상기 수술실별 스케줄 정보 및 상기 환자의 수술별 소요예정시간 정보 중 적어도 하나를 포함하는The number of treatment rooms in the hospital, the number of treatment rooms used by each medical staff of the hospital, outpatient support personnel information for each medical staff, session opening status information by day of the week for other medical staff in the same department of the hospital, bed information in the hospital room, patients in the hospital At least one of status information, status information for each department, type and number of operating rooms of the hospital, schedule information for each operating room, and expected time required for each operation of the patient
    병원의 의사결정 장치.Hospital decision-making equipment.
  6. 제 1 항에 있어서,The method of claim 1,
    상기 병원의 의사결정은,The hospital's decision-making
    상기 환자의 진료 스케줄 결정을 포함하고,determining the patient's care schedule;
    상기 프로세서는,The processor is
    상기 환자에 대한 진료정보 및 상기 예측된 환자의 의료 이벤트를 기초로 상기 환자의 진료 순위를 결정하고,Determining the medical care ranking of the patient based on the medical treatment information for the patient and the predicted patient medical event,
    상기 결정된 환자의 진료 순위 및 상기 병원의 의료진 정보를 더 고려하여 상기 환자의 진료 스케줄을 결정하는determining the patient's treatment schedule by further considering the determined patient's treatment order and medical staff information of the hospital
    병원의 의사결정 장치.Hospital decision-making equipment.
  7. 제 6 항에 있어서,7. The method of claim 6,
    상기 병원의 의료진 정보는,Medical staff information of the hospital,
    상기 의료진 별 진료 스케줄정보, 상기 환자 유형별 정보, 상기 환자의 진료에 의해 발생되는 시간정보, 상기 의료진 별 수술, 시술 및 연구 중 적어도 하나의 스케줄 정보, 상기 의료진 별 실적정보 및 상기 병원의 의료진 보직 여부 정보 중 적어도 하나를 포함하는The medical treatment schedule information for each medical staff, the information by the patient type, the time information generated by the patient's treatment, the schedule information of at least one of the surgery, the procedure, and the research by the medical staff, the performance information by the medical staff, and the position of the medical staff in the hospital at least one of the information
    병원의 의사결정 장치.Hospital decision-making equipment.
  8. 제 1 항에 있어서,The method of claim 1,
    상기 병원의 의사결정은,The hospital's decision-making
    상기 환자의 병상배정 결정을 포함하고,including determining the bed assignment of the patient;
    상기 프로세서는,The processor is
    상기 환자에 대한 진료정보 및 상기 예측된 환자의 의료 이벤트를 기초로 상기 환자의 병상배정 순위를 결정하고,Determining the bed allocation ranking of the patient based on the medical treatment information for the patient and the predicted patient's medical event,
    상기 결정된 환자의 병상 배정 순위를 더 고려하여 상기 환자의 병상배정을 결정하는determining the bed assignment of the patient by further considering the bed assignment order of the determined patient
    병원의 의사결정 장치.Hospital decision-making equipment.
  9. 복수의 환자에 대한 진료정보를 기 정해진 구조 형식의 데이터로 변환시키는 단계와,A step of converting the treatment information for a plurality of patients into data in a predetermined structural format;
    상기 기 정해진 구조 형식의 데이터를 입력으로, 상기 복수의 환자에 대해 발생될 수 있는 적어도 하나의 의료 이벤트를 정답으로하여 딥러닝 학습모델을 학습시키는 단계와,Learning a deep learning learning model by receiving the data in the predetermined structure format as an input and at least one medical event that may occur for the plurality of patients as an answer;
    환자에 대한 진료정보를 입력 받아, 상기 딥러닝 학습모델을 통해 상기 환자에 대한 적어도 하나의 의료 이벤트를 예측하는 단계와,receiving medical information about the patient and predicting at least one medical event for the patient through the deep learning learning model;
    상기 예측된 상기 환자에 대한 적어도 하나의 의료 이벤트, 상기 환자에 대한 진료정보 및 병원의 상태정보를 고려하여 상기 병원의 의사결정을 수행하는 단계를 포함하는Comprising the step of performing the decision-making of the hospital in consideration of the predicted at least one medical event for the patient, medical information about the patient, and the state information of the hospital
    병원의 의사결정 방법.How hospitals make decisions.
  10. 제 9 항에 있어서,10. The method of claim 9,
    상기 환자에 대한 진료정보는,Medical information about the patient,
    상기 환자가 상기 병원을 최초 방문한 시점에 따른 진료정보, 상기 환자가 상기 병원을 최초 방문한 시점 이후에 상기 병원을 방문한 시점에 따른 진료정보 및 상기 병원 이외의 타 병원에서 획득한 진료정보 중 적어도 하나를 포함하는At least one of medical treatment information according to the time when the patient first visited the hospital, medical information according to the time when the patient visited the hospital after the initial visit to the hospital, and medical information obtained from other hospitals other than the hospital containing
    병원의 의사결정 방법.How hospitals make decisions.
  11. 제 9 항에 있어서,10. The method of claim 9,
    상기 의료 이벤트는,The medical event is
    상기 환자의 퇴원 후 소정기간 이내 사망률, 상기 환자의 퇴원 후 소정기간 이내 재입원률, 상기 환자가 재입원한 시점에서부터 퇴원시점까지의 기간 및 상기 환자의 퇴원 후 소정기간 이내 심장병의 발병률 중 적어도 하나를 포함하는At least one of the mortality rate within a predetermined period after the patient's discharge, the readmission rate within a predetermined period after the patient's discharge, the period from the patient's readmission to the time of discharge, and the incidence rate of heart disease within the predetermined period after the patient's discharge containing
    병원의 의사결정 방법.How hospitals make decisions.
  12. 제 9 항에 있어서,10. The method of claim 9,
    상기 병원의 의사결정은,The hospital's decision-making
    상기 환자의 진료 스케줄 결정, 상기 환자의 병상배정 결정 및 상기 환자의 수술실 스케줄 결정을 포함하는Determining the patient's treatment schedule, determining the bed assignment of the patient, and determining the operating room schedule of the patient
    병원의 의사결정 방법.How hospitals make decisions.
  13. 제 9 항에 있어서,10. The method of claim 9,
    상기 병원의 상태정보는,The hospital's status information is
    상기 병원의 진료실 개수, 상기 병원의 의료진 별 사용한 진료실의 개수, 상기 의료진 별 외래 지원 인력정보, 상기 병원의 동일 진료과 내 타 의료진 요일별 세션 개설 상태정보, 상기 병원의 병실 침대정보, 상기 병원의 재원 환자 상태정보, 상기 진료과 별 상태정보, 상기 병원의 수술실의 종류와 개수, 상기 수술실별 스켈줄 정보 및 상기 환자의 수술별 소요예정시간 정보 중 적어도 하나를 포함하는The number of treatment rooms in the hospital, the number of treatment rooms used by each medical staff of the hospital, outpatient support personnel information for each medical staff, session opening status information by day of the week for other medical staff in the same department of the hospital, bed information in the hospital room, patients in the hospital At least one of status information, status information for each department, type and number of operating rooms of the hospital, schedule information for each operating room, and expected time required for each operation of the patient
    병원의 의사결정 방법.How hospitals make decisions.
  14. 제 9 항에 있어서,10. The method of claim 9,
    상기 병원의 의사결정은,The hospital's decision-making
    상기 환자의 진료 스케줄 결정을 포함하고,determining the patient's care schedule;
    상기 병원의 의사결정을 수행하는 단계는,The step of carrying out the decision-making of the hospital,
    상기 환자에 대한 진료정보 및 상기 예측된 환자의 의료 이벤트를 기초로 상기 환자의 진료 순위를 결정하고,Determining the medical care ranking of the patient based on the medical treatment information for the patient and the predicted patient medical event,
    상기 결정된 환자의 진료 순위 및 상기 병원의 의료진 정보를 더 고려하여 상기 환자의 진료 스케줄을 결정하는determining the patient's treatment schedule by further considering the determined patient's treatment order and medical staff information of the hospital
    병원의 의사결정 방법.How hospitals make decisions.
  15. 제 14 항에 있어서,15. The method of claim 14,
    상기 병원의 의료진 정보는,Medical staff information of the hospital,
    상기 의료진 별 진료 스케줄정보, 상기 환자 유형별 정보, 상기 환자의 진료에 의해 발생되는 시간정보, 상기 의료진 별 수술, 시술 및 연구 중 적어도 하나의 스케줄 정보, 상기 의료진 별 실적정보 및 상기 병원의 의료진 보직 여부 정보 중 적어도 하나를 포함하는The medical treatment schedule information for each medical staff, the information by the patient type, the time information generated by the patient's treatment, the schedule information of at least one of the surgery, the procedure, and the research by the medical staff, the performance information by the medical staff, and the position of the medical staff in the hospital at least one of the information
    병원의 의사결정 방법.How hospitals make decisions.
  16. 제 9 항에 있어서,10. The method of claim 9,
    상기 병원의 의사결정은,The hospital's decision-making
    상기 환자의 병상배정 결정을 포함하고,including determining the bed assignment of the patient;
    상기 병원의 의사결정을 수행하는 단계는,The step of carrying out the decision-making of the hospital,
    상기 환자에 대한 진료정보 및 상기 예측된 환자의 의료 이벤트를 기초로 상기 환자의 병상배정 순위를 결정하고,Determining the bed allocation ranking of the patient based on the medical treatment information for the patient and the predicted patient's medical event,
    상기 결정된 환자의 병상 배정 순위를 더 고려하여 상기 환자의 병상배정을 결정하는determining the bed assignment of the patient by further considering the bed assignment order of the determined patient
    병원의 의사결정 방법.How hospitals make decisions.
  17. 컴퓨터 프로그램을 저장하고 있는 컴퓨터 판독 가능 기록매체로서,As a computer-readable recording medium storing a computer program,
    상기 컴퓨터 프로그램은, 프로세서에 의해 실행되면,The computer program, when executed by a processor,
    복수의 환자에 대한 진료정보를 기 정해진 구조 형식의 데이터로 변환시키는 단계와,A step of converting the treatment information for a plurality of patients into data in a predetermined structural format;
    상기 기 정해진 구조 형식의 데이터를 입력으로, 상기 복수의 환자에 대해 발생될 수 있는 적어도 하나의 의료 이벤트를 정답으로하여 딥러닝 학습모델을 학습시키는 단계와,Learning a deep learning learning model by receiving the data in the predetermined structure format as an input and at least one medical event that may occur for the plurality of patients as an answer;
    환자에 대한 진료정보를 입력 받아, 상기 딥러닝 학습모델을 통해 상기 환자에 대한 적어도 하나의 의료 이벤트를 예측하는 단계와,receiving medical information about the patient and predicting at least one medical event for the patient through the deep learning learning model;
    상기 예측된 상기 환자에 대한 적어도 하나의 의료 이벤트, 상기 환자에 대한 진료정보 및 병원의 상태정보를 고려하여 상기 병원의 의사결정을 수행하는 단계를 포함하는 방법을 상기 프로세서가 수행하도록 하기 위한 명령어를 포함하는Instructions for causing the processor to perform a method comprising the step of performing decision-making of the hospital in consideration of the predicted at least one medical event for the patient, medical treatment information for the patient, and state information of the hospital containing
    컴퓨터 판독 가능한 기록매체.computer readable recording medium.
PCT/KR2021/008298 2020-07-21 2021-06-30 Apparatus, method, and computer-readable recording medium for decision-making in hospital WO2022019514A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
KR10-2020-0090627 2020-07-21
KR1020200090627A KR102501195B1 (en) 2020-07-21 2020-07-21 Apparatus, method, computer-readable storage medium and computer program for decision making by hospital

Publications (1)

Publication Number Publication Date
WO2022019514A1 true WO2022019514A1 (en) 2022-01-27

Family

ID=79729277

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/KR2021/008298 WO2022019514A1 (en) 2020-07-21 2021-06-30 Apparatus, method, and computer-readable recording medium for decision-making in hospital

Country Status (2)

Country Link
KR (1) KR102501195B1 (en)
WO (1) WO2022019514A1 (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116052887A (en) * 2023-03-01 2023-05-02 联仁健康医疗大数据科技股份有限公司 Method and device for detecting excessive inspection, electronic equipment and storage medium

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101261177B1 (en) * 2011-11-29 2013-05-09 삼성에스디에스 주식회사 Clinical decision support system and method
KR20170140757A (en) * 2016-06-10 2017-12-21 한국전자통신연구원 A clinical decision support ensemble system and the clinical decision support method by using the same
JP2020003862A (en) * 2018-06-25 2020-01-09 清水建設株式会社 Schedule generation device, medical facility management system, program, schedule generation method, medical facility management method, and learning method of artificial intelligence
KR20200003407A (en) * 2017-07-28 2020-01-09 구글 엘엘씨 Systems and methods for predicting and summarizing medical events from electronic health records
US20200066397A1 (en) * 2018-08-23 2020-02-27 General Electric Company Multifactorical, machine-learning based prioritization framework for optimizing patient placement

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20180026561A (en) 2009-01-19 2018-03-12 비오메리으 Methods for determining the likelihood of a patient contracting a nosocomial infection and for determining the prognosis of the course of a septic syndrome

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101261177B1 (en) * 2011-11-29 2013-05-09 삼성에스디에스 주식회사 Clinical decision support system and method
KR20170140757A (en) * 2016-06-10 2017-12-21 한국전자통신연구원 A clinical decision support ensemble system and the clinical decision support method by using the same
KR20200003407A (en) * 2017-07-28 2020-01-09 구글 엘엘씨 Systems and methods for predicting and summarizing medical events from electronic health records
JP2020003862A (en) * 2018-06-25 2020-01-09 清水建設株式会社 Schedule generation device, medical facility management system, program, schedule generation method, medical facility management method, and learning method of artificial intelligence
US20200066397A1 (en) * 2018-08-23 2020-02-27 General Electric Company Multifactorical, machine-learning based prioritization framework for optimizing patient placement

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116052887A (en) * 2023-03-01 2023-05-02 联仁健康医疗大数据科技股份有限公司 Method and device for detecting excessive inspection, electronic equipment and storage medium
CN116052887B (en) * 2023-03-01 2023-06-27 联仁健康医疗大数据科技股份有限公司 Method and device for detecting excessive inspection, electronic equipment and storage medium

Also Published As

Publication number Publication date
KR20220011549A (en) 2022-01-28
KR102501195B1 (en) 2023-02-21

Similar Documents

Publication Publication Date Title
US20050075904A1 (en) System and method for automatically generating evidence-based assignment of care providers to patients
WO2010067981A9 (en) System and method for interfacing medical device information
WO2022145782A2 (en) Big data and cloud system-based artificial intelligence emergency medical care decision making and emergency patient transporting system and method therefor
US20070168225A1 (en) Workflow generator for medical-clinical facilities
US20070038474A1 (en) Workflow and communications logging functions of an automated medical case management system
WO2020080819A1 (en) Oral health prediction apparatus and method using machine learning algorithm
WO2017183878A1 (en) Method and system for providing medical service, and non-transitory computer-readable recording medium
WO2022019514A1 (en) Apparatus, method, and computer-readable recording medium for decision-making in hospital
US20180226141A1 (en) Patient coordination system and method
WO2016179885A1 (en) User terminal, expert platform and expert consultation request method based on user terminal
WO2018105995A2 (en) Device and method for health information prediction using big data
WO2020032562A2 (en) Bioimage diagnosis system, bioimage diagnosis method, and terminal for executing same
KR101729143B1 (en) System for managing medical process
WO2022260293A1 (en) Method for vectorizing medical data for machine learning, and data conversion device and data conversion program in which same is implemented
US20160239616A1 (en) Medical support system, method and apparatus for medical care
JP7388356B2 (en) Medical information processing system, medical information processing device, and medical information processing method
CN111798951A (en) Self-help medical guide system for infectious diseases
WO2020022825A1 (en) Method and electronic device for artificial intelligence (ai)-based assistive health sensing in internet of things network
WO2023121051A1 (en) Patient information provision method, patient information provision apparatus, and computer-readable recording medium
WO2019164273A1 (en) Method and device for predicting surgery time on basis of surgery image
Monks et al. Using simulation to help hospitals reduce emergency department waiting times: Examples and impact
WO2022225122A1 (en) Patient-customized electronic medical questionnaire system, medical questionnaire terminal, and medical questionnaire method using same
CN106169029A (en) A kind of communication system being applied to medical field and method
CN103876836B (en) Check queue control method and its control device
WO2021040200A1 (en) Data transmission system and data transmission method for transmitting/receiving medical information data

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21846242

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 21/06/2023)

122 Ep: pct application non-entry in european phase

Ref document number: 21846242

Country of ref document: EP

Kind code of ref document: A1