JP2020003862A - Schedule generation device, medical facility management system, program, schedule generation method, medical facility management method, and learning method of artificial intelligence - Google Patents

Abstract

To provide a schedule generation device which can easily generate a schedule of a medical facility.SOLUTION: In a medical facility management system 1, a schedule generation device 20 comprises: an acquisition unit 21 which acquires operation list data; a storage unit 22 which stores a learned model; and a generation unit 23 which generates medical facility schedule data. The acquisition unit 21 stores the acquired operation list data and the learned model in the storage unit 22. The generation unit 23 generates medical facility schedule data which becomes the schedule during a predetermined period in the medical facility, on the basis of the operation list data and the learned model stored in the storage unit 22.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a schedule generation device, a medical facility management system, a program, a schedule generation method, a medical facility management method, and a method of learning artificial intelligence.

  In a medical facility such as a hospital, medical practices such as medical treatment in a doctor's office and surgery in an operating room are performed. Since these facilities are used sequentially by medical personnel such as a plurality of doctors, a schedule of medical facilities is created. Conventionally, a schedule of medical equipment has been manually created by a nurse, for example, at the beginning of a week or the beginning of a month. As a technique for creating a schedule, for example, there is a technique disclosed in Patent Document 1.

JP 2012-103810 A

  However, a number of factors need to be considered when creating a schedule for a medical facility. For this reason, when creating a schedule for medical equipment, it is difficult to create a suitable schedule unless the nurse is an experienced nurse. Further, even experienced nurses have a problem that it takes time to create a suitable schedule.

  The problem to be solved by the present invention is to provide a schedule generation device, a medical facility management system, a program, a schedule generation method, a medical facility management method, and an artificial intelligence learning method that can easily generate a schedule of medical facilities. .

  One aspect of the present invention for solving the above-described problem is an acquisition unit that acquires medical-related act data related to a medical-related act performed in a medical facility, a medical-related act performed in a medical facility, and the medical facility A generation unit that generates a use schedule of the medical facility by using a learned model that has learned a relationship with schedule related information related to the use schedule of the medical facility.

  Further, one aspect of the present invention is a learning unit that updates a learned model that has learned a relationship between a medical-related action performed in a medical facility and schedule-related information related to a use schedule of the medical facility. And a schedule generation device that generates a use schedule of the medical facility by using the learned model updated by the section.

  One embodiment of the present invention is a program for causing a computer to operate as a schedule generating device, and for causing the computer to function as each unit included in the schedule generating device.

  One embodiment of the present invention is a program for operating a computer as a medical facility management system, and is a program for causing the computer to function as each unit included in the medical facility management system.

  Further, one aspect of the present invention is a schedule generation method in a schedule generation device, wherein the medical-related activity data related to a medical-related activity performed in a medical facility is acquired, and the medical-related activity performed in a medical facility, This is a schedule generation method for generating a use schedule of the medical facility by using a learned model that has learned a relationship with schedule related information related to a use schedule of the medical facility.

  One embodiment of the present invention is a medical facility management method in a medical facility management system, wherein the learning unit performs a medical-related action performed in the medical facility, and schedule-related information related to a use schedule of the medical facility. Is a medical facility management method in which the learned model that has learned the relationship is updated, and the schedule generation device generates a use schedule of the medical facility by using the learned model updated by the learning unit.

  In addition, one embodiment of the present invention associates medical-related activity data relating to medical-related activities performed in a medical facility with schedule-related information related to the use schedule of the medical facility, and causes the artificial intelligence to learn the new information, thereby enabling new learning. This is an artificial intelligence learning method for constructing a learned model for generating schedule-related information from medical-related action data.

  As described above, according to the present invention, it is possible to provide a schedule generation device, a medical facility management system, a program, a schedule generation method, a medical facility management method, and a learning method of artificial intelligence that can easily generate a schedule of a medical facility. Can be.

1 is a schematic block diagram illustrating a configuration of a medical facility management system according to an embodiment of the present invention. It is a figure showing an example of operation list data. It is a figure showing a learned model. It is a figure showing an example of a use schedule of a medical facility. It is a table | surface explaining the content of the use schedule of a medical facility. 6 is a flowchart illustrating an example of a process in the schedule generation device. It is a flow chart which shows an example of processing in a learning device.

  Hereinafter, a medical facility management system according to an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a schematic block diagram showing the configuration of a medical facility management system 1 according to one embodiment of the present invention. As shown in FIG. 1, the medical facility management system 1 includes an input device 10, a schedule generation device 20, a learning device 30, and an output device 40.

  The input device 10 is configured by a device that can be operated by an operator, such as a keyboard and a touch panel. Various information can be input to the input device 10 by an operation of an operator who inputs various data. The input device 10 transmits to the schedule generation device 20 and the learning device 30 various input data, for example, operation list data (medical-related activity data) regarding a hospital (operation (medical activity) performed in a medical facility).

  As shown in FIG. 2, the surgery list data is data on a patient who is scheduled to undergo surgery for a certain period of time, in this case, one day (surgery-scheduled patient). The items of the operation list data include, for example, a patient's name, a medical department, and an operation method. The operation list data may include other data regarding the patient, such as age, medical history, and the number of past operations.

  The schedule generation device 20 includes an arithmetic device such as a CPU (Central Processing Unit) and a storage device such as a memory. The schedule generation device 20 functionally includes an acquisition unit 21, a storage unit 22, and a generation unit 23. The acquisition unit 21 receives and acquires the operation list data transmitted from the input device 10 and the learned model transmitted from the learning device 30. The acquisition unit 21 causes the storage unit 22 to store the acquired surgery list data and the learned model. The learned model is a model (data) for generating medical facility schedule data from new operation list data. Note that the learned model may be a model other than the learned model constructed by the learning device 30 or may be a model stored in the storage unit 22 in advance.

  The generation unit 23 generates medical facility schedule data that is a predetermined period (scheduling period) of the medical facility, for example, a one-day schedule, based on the operation list data and the learned model stored in the storage unit 22. FIG. 3 is a diagram showing an example of the medical facility schedule data, and FIG. 4 is a table for explaining the contents of the medical facility use schedule data. As shown in FIG. 3, the medical facility schedule data includes a daily schedule of 10 operating rooms of operating room A to operating room I and operating room X. The medical facility schedule data may be one day, one week or one month. The medical facility included in the medical facility schedule data may be a facility other than the operating room, for example, an examination room, a treatment room, an examination room, and the like.

  The medical facility schedule data includes medical departments, surgical procedures, medical personnel, such as surgeons, nurses (three), and anesthesiologists (a staffing chart at the time of operation) in each operation performed in the operating room. Schedule related information related to the medical facility schedule is included. In addition, the items to be tabulated as a total include (totals No. 1 to No. 10), various items (predetermined cumulative values) such as labor costs (total cost), working hours (total working hours), and profits (total sales amount). Have been.

  The learning device 30 includes an arithmetic device such as a CPU and a storage device such as a memory. As shown in FIG. 1, the learning device 30 functionally includes an acquisition unit 31, an analysis unit 32, a storage unit 33, and a learning unit 34. The acquisition unit 31 receives and acquires the operation list data transmitted from the input device and the medical facility schedule data transmitted from the schedule generation device 20. The acquisition unit 31 outputs the acquired operation list data to the analysis unit 32. The acquiring unit 31 causes the storage unit 33 to store the acquired medical facility schedule data.

  The analysis unit 32 analyzes the operation list data output from the acquisition unit 31 and classifies the operation list data into data on medical characteristics of medical personnel, characteristics of medical facilities, and usage modes of medical facilities. Medical personnel data is data on medical characteristics of medical personnel, medical facility data is data on characteristics of medical facilities, operative data is data on surgical procedures, and data on the number of frames is data on the number of frames in medical facilities. Data. The analysis unit 32 causes the storage unit 33 to store the classified medical personnel data, medical facility data, surgical procedure data, and frame number data. FIG. 5A is a diagram showing operation-related data, FIG. 5B is a diagram showing medical facility data, FIG. 5C is a diagram showing operative data, and FIG. 5D is a diagram showing frame number data. It is.

  As shown in FIG. 5A, the medical personnel data includes the number of doctors performing the surgery and various properties of each doctor. Further, the medical personnel data includes the number of nurses assisting the surgery and various properties of each nurse, the number of anesthesiologists performing anesthesia at the time of surgery, and various properties of each anesthesiologist. Furthermore, the medical personnel data also includes various properties of the patient undergoing the operation.

  The physician properties include various data such as specialized departments, years of experience, relationships with nurses and anesthesiologists, working hours, holidays, and hours. The specialized course is a department specialized by the doctor, for example, internal medicine, orthopedic surgery, respiratory medicine, and the like. Years of experience is the number of years the physician has spent as a physician. In addition, the number of years of experience after the physician starts performing surgery may be used.

  The relationship with the nurse is, for example, compatibility with the nurse, and is quantified as, for example, the number of times the nurse and the surgeon performed surgery together. Regarding the relationship with the nurse, the relationship with all nurses may be taken into consideration, or the relationship may be limited to nurses that have a characteristic as compatibility, for example, nurses with particularly good or bad compatibility. The relationship with an anesthesiologist is similar to the relationship with a nurse.

  The upper limit of working hours is the upper limit of working hours defined as the working hours of each doctor. The upper limit of working hours may be determined for each doctor or may be a common time. A vacation day is a day set as a vacation of each doctor. The time unit is a time unit in which each doctor can work continuously.

  Nurse and anesthesiologist properties are similar to physician properties. The patient properties include various data such as age, history, and number of times. The age is the age of the patient, and the history is the history of the patient. The number of times is the number of operations that the patient has undergone in the past.

  As shown in FIG. 5B, the medical facility data includes the number of operating rooms and various properties of each operating room. The table also includes the number of rooms in the preliminary operating room, various properties of each preliminary room, the number of surgical tools (use tools) (the number of surgical tools), and various properties of each surgical tool. The properties of the operating room include the size (area) of each operating room, medical facilities, facility facilities, and the presence or absence of the front and rear rooms.

  As shown in FIG. 5 (C), the surgical procedure data includes various properties of resources and unit sales price. The resource includes properties such as “person”, “thing”, “medical facility”, and “operating room size”. As shown in FIG. 5D, the frame number data includes various properties such as the number of frames per week, the number of frames per day, and the frame time.

  The learning unit 34 makes AI (Artificial Intelligence) learn by associating the operation list data with schedule-related information related to the medical facility schedule. The learning unit 34 uses the medical personnel data, the medical facility data, the operative method data, and the data on the number of frames (hereinafter referred to as “analysis data”) stored in the storage unit 33 based on the learning algorithm and the analysis data. Learn the relationship between medical facility schedule data and update and build the learned model. For example, the learning unit 34 analyzes each item included in the analysis data and acquires a feature amount of each item. The learning unit 34 learns the relationship between the feature amount and the medical facility schedule data, and generates a learned model. Here, the learning in the learning unit 34 may learn, for example, a condition indicating the relationship between the feature amount and the medical facility schedule data. The condition in the learning unit 34 is a model structure representing the correlation between the feature amount of the analysis data and the medical facility schedule data, or various parameters for determining the model structure.

  The learning of the learning unit 34 may be any learning method, for example, using any one of machine learning using AI technology, reinforcement learning, and deep learning using a neural network including a plurality of hidden layers. Can be. The learning method may be supervised learning or unsupervised learning.

  For example, at the start of learning, the learning unit 34 does not know the correlation between the feature amount of the analysis data and the medical facility schedule data, but interprets the correlation by gradually identifying features as the learning proceeds. I do. As the learning progresses, the learning unit 34 can make the condition indicating the correlation between the feature amount of the analysis data and the medical facility schedule data closer to the optimal solution. The learning unit 34 learns the relationship between the feature amount of the analysis data and the medical facility schedule data. However, the learning unit 34 determines the relationship between the analysis data and the medical facility schedule data without using the feature amount of the analysis data. The conditions to be represented may be learned. In this case, the learning unit 34 may learn, for example, a condition indicating a relationship between each item included in the analysis data and a predetermined item of the medical facility schedule data.

  The learning unit 34 performs a process of giving a label to the medical facility schedule data. Here, when the learning method performed by the learning unit 34 is supervised learning, a process of assigning a label as teacher data to the medical facility schedule data is performed. If the learning method performed by the learning unit 34 is unsupervised learning, labeling does not need to be performed, and in that case, the labeling function does not necessarily have to be provided.

  The label given to the medical facility schedule data may be for a predetermined item of the medical facility schedule data. For example, it may be added to the profit or working hours in the medical facility schedule data shown in FIG. Alternatively, it may be provided when the profit and the working time have a predetermined relationship (when the working time does not exceed the predetermined time and the profit is equal to or more than a predetermined amount). Labels may be given to other items.

  Next, processing in the medical facility management system 1 will be described. In the medical facility management system 1, the schedule generation device 20 schedules an operation in an operating room provided in a medical facility based on the operation list data input to the input device 10 and the trained model learned by the learning device 30. Is generated. In the medical facility management system 1, the learning device 30 updates the learned model based on the relationship between the operation list data input to the input device 10 and the medical facility schedule data generated by the schedule generation device 20.

  In generating the medical facility schedule data and updating the learned model, the medical facility management system 1 firstly inputs, from the input device 10, the operation list data relating to the operation scheduled for a certain period, for example, during a week of a predetermined week. Is entered. The operation list data is input, for example, by a nurse serving as an operator. The operation list data input to the input device 10 is transmitted to the schedule generation device 20 and the learning device 30. The schedule generation device 20 generates medical facility schedule data. The schedule generation device 20 transmits the generated medical facility schedule data to the learning device 30, and the learning device 30 updates the learned model using the transmitted medical facility schedule data. Hereinafter, each of the process of generating medical facility schedule data by the schedule generation device 20 and the process of updating the learned model by the learning device 30 will be described.

  FIG. 6 is a flowchart illustrating an example of processing in the schedule generation device. The schedule generation device generates medical facility schedule data. As shown in FIG. 6, in the schedule generation device 20, first, the acquisition unit 21 receives the operation list data transmitted from the input device 10 (step S101). Here, the operation list data acquired from the input device 10 is data (new medical-related activity data) for which medical facility schedule data is to be generated. Next, the schedule generation device 20 causes the generation unit 23 to read the learned model stored in the storage unit 22 (Step S102). Note that when the acquisition unit 21 receives the learned model (updated learned model) transmitted from the learning device 30, the acquisition unit 21 updates the learned model stored in the storage unit 22 as needed. Update and store.

  Subsequently, the schedule generation device 20 causes the generation unit 23 to generate medical facility schedule data (step S103). The generation unit 23 generates medical facility schedule data based on the operation list data and the learned model stored in the storage unit 22.

  Thereafter, the schedule generating device 20 causes the output device 40 to output the generated medical facility schedule data (step S104). Further, the schedule generation device 20 transmits the generated medical facility schedule data to the learning device 30 (Step S105). Thus, the process of the schedule generation device 20 ends.

  FIG. 7 is a flowchart illustrating an example of a process in the learning device. In the learning device 30, a learned model is constructed (updated). As shown in FIG. 7, the learning device 30 first receives the operation list data transmitted from the input device 10 in the acquisition unit 31 (step S201). In the learning device 30, the acquisition unit 31 receives the medical facility schedule data transmitted from the schedule generation device 20 (step S202). The acquisition unit 31 outputs the received operation list data to the analysis unit 32, and stores the medical facility schedule data in the storage unit 33.

  Subsequently, the learning device 30 analyzes the operation list data transmitted from the input device 10 in the analysis unit 32 (step S203), and converts the operation list data into analysis data (medical personnel data, medical facility data, operative method data). , And frame number data) and store them in the storage unit 33 (step S204).

  Next, in the learning device 30, the learning unit 34 reads the medical facility schedule data, labels the medical facility schedule data (step S205), and stores the data in the storage unit 33. When the labeled medical facility schedule data is accumulated to some extent, the learning device 30 reads out the analysis data and the labeled medical facility schedule data in the learning unit 34, and performs learning based on the learning algorithm (step S1). S206). The learning device 30 generates a learned model by performing learning, and constructs the learned model in the generated learned model (step S207). Thereafter, the updated learned model is transmitted to the schedule generation device 20 (step S208), and the process of the learning device 30 ends.

  When constructing the learned model, an object for optimization or the like may be determined for each object of the schedule. For example, in the case of a daily schedule, optimization of the total number of slots, optimization of the order of slots, optimization of the number of slots for each medical department, and maximization of sales amount may be performed. When the schedule is a weekly schedule, the total number of slots, the number of slots for each medical department, and the sales amount may be maximized. Further, when targeting persons such as doctors and nurses, the total working hours of individuals may be minimized (leveled) and the total labor cost of individuals may be minimized.

  In the above-described embodiment, the learning device 30 learns the medical facility schedule data to generate a learned model, and the schedule generation device 20 generates a medical facility schedule using the medical-related data and the learned model. are doing. Therefore, appropriate medical facility schedule data can be generated based on the medical facility schedule data generated in the past. Therefore, a schedule for medical equipment can be easily generated.

  In the above-described embodiment, a case has been described in which the schedule generation device 20 and the learning device 30 are different devices. However, the function of the schedule generation device 20 may be provided in the learning device 30. That is, the schedule generation device 20 may store the learned model and generate the medical facility schedule without using the learning device 30. The learning device 30 may be individually configured as a device having a learning function and a function of generating medical facility schedule data based on a learned model.

  Further, in the above-described embodiment, the medical-related act is an operation, and the medical facility is an operating room. However, another medical-related act or a medical facility may be used. For example, medical-related actions may include medical examinations, examinations, and treatments other than surgery. Further, the medical facility may be a medical examination room, a treatment room, an examination room, or the like.

  The schedule generation device 20 or the learning device 30 in the above-described embodiment may be realized by a computer. In that case, a program for realizing this function may be recorded on a computer-readable recording medium, and the program recorded on this recording medium may be read and executed by a computer system. Here, the “computer system” includes an OS and hardware such as peripheral devices. The “computer-readable recording medium” refers to a portable medium such as a flexible disk, a magneto-optical disk, a ROM, and a CD-ROM, and a storage device such as a hard disk built in a computer system. Further, a “computer-readable recording medium” refers to a communication line for transmitting a program via a network such as the Internet or a communication line such as a telephone line, which dynamically holds the program for a short time. Such a program may include a program that holds a program for a certain period of time, such as a volatile memory in a computer system serving as a server or a client in that case. The program may be for realizing a part of the functions described above, or may be a program that can realize the functions described above in combination with a program already recorded in a computer system, It may be realized using a programmable logic device such as an FPGA (Field Programmable Gate Array).

  As described above, the embodiments of the present invention have been described in detail with reference to the drawings. However, the specific configuration is not limited to the embodiments, and includes a design and the like within a range not departing from the gist of the present invention.

Reference Signs List 1 medical facility management system 10 input device 20 schedule generation device 21 acquisition unit 22 storage unit 23 generation unit 30 learning device 31 acquisition unit 32 analysis unit 33 storage unit 34 learning unit 40 output device

Claims (9)

An acquisition unit that acquires medical-related act data related to medical-related acts performed in a medical facility;
A generation unit that generates a use schedule of the medical facility, using a learned model that has learned a relationship between a medical action performed in the medical facility and schedule-related information related to the use schedule of the medical facility,
A schedule generation device comprising:   2. The schedule generation device according to claim 1, wherein the medical-related action data is at least one of a list of patients to be operated on and a surgical procedure during a predetermined scheduling period, medical personnel, and data on medical facilities. 3.   3. The schedule generation device according to claim 1, wherein the schedule-related information is at least one of a schedule of a medical facility, a schedule of a patient operation, a staffing table at the time of operation, and a predetermined total value of the operation. . A learning unit that updates a learned model that has learned the relationship between medical-related actions performed in a medical facility and schedule-related information related to the use schedule of the medical facility,
A schedule generation device that generates a use schedule of the medical facility by using the learned model updated by the learning unit;
Medical facility management system equipped with.   A program for causing a computer to operate as the schedule generation device according to any one of claims 1 to 3, wherein the program causes the computer to function as each unit included in the schedule generation device.   A program for operating a computer as the medical facility management system according to claim 4, wherein the program causes the computer to function as each unit included in the medical facility management system. A schedule generation method in a schedule generation device,
Get medical activity data related to medical activities performed in medical facilities,
A schedule generation method that generates a use schedule of the medical facility by using a learned model that has learned a relationship between a medical action performed in the medical facility and schedule-related information related to the use schedule of the medical facility. A medical facility management method in a medical facility management system,
The learning unit updates the learned model that has learned the relationship between the medical-related action performed in the medical facility and the schedule-related information related to the use schedule of the medical facility,
A medical facility management method, wherein a schedule generation device generates a use schedule of the medical facility using the learned model updated by the learning unit. Medical-related act data related to medical-related acts performed in the medical facility, and schedule-related information related to the use schedule of the medical facility, and let artificial intelligence learn by associating,
A method of learning artificial intelligence that builds a trained model that generates schedule-related information from new medical activity data.

Images