Information processing device

Abstract

We provide a device and method for predicting an individual's medical expenses based on information such as daily behavior, habits, and test results. [Solution] The information processing device has a learning data collection means for acquiring personal health records and medical expenses, which are data related to an individual's health and medical care, from an external device; a learning dataset generation means for acquiring from a user a combination of external device identification information that identifies the external device and user identification information that identifies the individual, and linking the collected personal health records and medical expenses based on the acquired combination and integrating them as a learning dataset; a model generation means for generating a machine learning model that outputs a predicted value of medical expenses when a personal health record is input by training the generated learning dataset into a machine learning model; and a medical expense prediction means for inputting a personal health record for a single user into the trained machine learning model and calculating a predicted value of medical expenses for the single user.

Description

[Technical Field]

[0001] This project relates to technology for predicting individual medical expenses and technology for encouraging insured persons to take actions that contribute to reducing medical expenses. [Background technology]

[0002] Some people rarely visit medical institutions and therefore their medical expenses are lower than the health insurance premiums they pay, while others pay more in medical expenses than their health insurance premiums. Incidentally, people who have low medical expenses may have certain characteristics in their daily behavior, habits, or health checkup results. Summary of the Invention [Problem to be solved by the invention]

[0003] In this situation, if it were possible to predict each individual's future medical expenses, it would be possible to increase the predictability of health insurance expenditures, thereby potentially increasing the reliability of health insurance management. Furthermore, if it were possible to encourage other insured persons to adopt the same tendency as those with relatively low medical expenses, it could potentially lead to a reduction in overall medical expenses under health insurance.

[0004] Therefore, based on the above viewpoint, the present invention aims to provide an information processing device that predicts the medical expenses of an individual based on information such as daily behavior, habits, and test results. [Means for solving the problem]

[0005] One form of the information processing device disclosed includes: a training data collection means for acquiring from an external device a personal health record and medical expenses, which are element data of a training dataset to be trained by a machine learning model, the personal health record and medical expenses being data related to an individual's health and medical care; a training dataset generation means for acquiring from a user a combination of external device identification information that identifies the external device and user identification information that identifies the individual, and linking the collected personal health record and medical expenses based on the acquired combination to integrate them as the training dataset; a model generation means for generating the machine learning model that outputs a predicted value of the medical expenses when the personal health record is input by having the machine learning model learn the generated training dataset; and a medical expense prediction means for inputting the personal health record of one of the users into the trained machine learning model and calculating a predicted value of the medical expenses for the one user. a correction information storage means for storing a type of health insurance and a correction coefficient defined for each type of health insurance in association with each other; an average medical expense storage means for storing, for each age group of insured persons in the health insurance, an average medical expense of the insured persons belonging to the age group; a first coefficient specifying means for acquiring information on the type of health insurance to which the one user is enrolled and extracting from the correction information storage means a first coefficient which is the correction coefficient corresponding to the acquired information on the type of health insurance; and a second coefficient specifying means for specifying, based on the personal health record of the one user, a first coefficient which increases as the one user engages in more health promotion activities. a second coefficient specifying means for calculating two coefficients; a third coefficient specifying means for calculating a third coefficient that increases as the predicted value of the medical expenses for the user is smaller than the average medical expenses for the age group to which the user belongs; a refund amount calculating means for calculating the amount of electronic currency to be refunded to the user by correcting the base amount set by the insurer of the health insurance to which the user belongs based on the first coefficient, the second coefficient, and the third coefficient; and a refund implementation means for adding the calculated amount of electronic currency to be refunded to the user to the balance of the electronic currency managed by the user. The present invention is characterized by having the following. [Effects of the Invention]

[0006] The disclosed information processing device predicts an individual's medical expenses based on information such as daily behavior, habits, and test results. [Brief explanation of the drawings]

[0007] [Figure 1] 1 is a diagram illustrating an overview of an information processing device according to an embodiment of the present invention. [Figure 2] 1 is a functional block diagram of an information processing device according to an embodiment of the present invention; [Figure 3] FIG. 1 is a diagram illustrating an example of a hardware configuration of an information processing device according to an embodiment of the present invention. [Figure 4] 10 is a flowchart illustrating an example of a flow of a medical cost prediction process performed by the information processing device according to the present embodiment. [Figure 5] 10 is a flowchart illustrating an example of a flow of a point redemption process performed by the information processing device according to the present embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0008] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described with reference to the drawings. (Operation principle of the information processing device according to this embodiment)

[0009] The operating principle of an information processing device (hereinafter simply referred to as "the device") 100 according to this embodiment will be described with reference to Figures 1 and 2. Figure 1 is a diagram showing the connection relationship between the device 100 and another device 380, and Figure 2 is a functional block diagram of the device 100.

[0010] 1, the device 100 is connected to an external device 380 via a communication network 490. The communication network 490 may be wired or wireless. The external device 380 may include a wearable terminal worn by the user 400, a mobile information terminal operated by the user 400, a server managed by a health insurance association 230 to which the user 400 belongs, a server managed by a medical institution used by the user 400, an information provision record disclosure system (so-called My Number Portal), which is an online service operated by the government, and the like.

[0011] As shown in Figure 2, the present device 100 has a user identification information storage means 110, a model parameter storage means 120, a correction information storage means 130, an average medical expense storage means 140, a training data collection means 150, a training data set generation means 160, a model generation means 170, a medical expense prediction means 180, a first coefficient determination means 190, a second coefficient determination means 200, a third coefficient determination means 210, a refund amount calculation means 220, and a refund implementation means 230.

[0012] The user identification information storage means 110 stores, in association with each other, external device identification information 390 that identifies an external device 380 connected via a communication network 490 and user identification information 410 that identifies an individual 400. The user identification information storage means 110 is a database for handling element data 330, which will be described later, in an integrated manner. The model parameter storage means 120 stores parameters that define the operation of the trained machine learning model 310.

[0013] The correction information storage means 130 stores health insurance types 420 in association with correction coefficients 430 defined for each health insurance type 420. The health insurance types 420 may be classified as national health insurance or employee health insurance, or classified by region, health insurance association, or various mutual aid associations. The average medical expenses storage means 140 stores, for each age group of insured persons under health insurance, the average medical expenses 370 of insured persons belonging to each age group.

[0014] The training data collection means 150 acquires, from an external device 380, a personal health record 340 and medical expenses 350, which are element data 330 of a training dataset 320 used to train a machine learning model 310 and are data related to the health and medical care of an individual 400. The personal health record 340 and medical expenses 350 acquired by the training data collection means 150 include external device identification information 390 that identifies the external device 380 from which the data was acquired, and user identification information 410 that indicates to whom each piece of data belongs.

[0015] Personal health record 340 includes information on an individual's exercise status and physical condition obtained from wearable devices and mobile information devices worn or operated by the user, health checkup results and medical examination history obtained from the information provision record disclosure system (so-called My Number Portal), an online service operated by the government, and servers managed by health insurance associations and medical institutions. (Past) medical expenses 350 are obtained from the information provision record disclosure system and servers managed by health insurance associations and medical institutions.

[0016] The training dataset generation means 160 acquires from the user 400 a combination of external device identification information 390 that identifies the external device 380 and user identification information 410 that identifies the individual 400. Then, based on the acquired combination, the training dataset generation means 160 links the collected personal health records 340 (330) and medical expenses 350 (330) and integrates them as a training dataset 320.

[0017] The model generation means 170 generates a machine learning model 310 that outputs a predicted value 360 ​​of medical expenses 350 when a personal health record 340 is input, by having the machine learning model 310 learn the generated learning dataset 320. The learning algorithm of the machine learning model 310 is not particularly limited.

[0018] The medical cost prediction means 180 inputs a personal health record 340 for one user 400 into the trained machine learning model 310 and calculates a predicted value 360 ​​of medical costs 350 for one user 400.

[0019] The first coefficient identification means 190 acquires information about the type 420 of health insurance to which one user 400 is enrolled. Then, the first coefficient identification means 190 extracts, from the correction information storage means 130, the first coefficient 440, which is the correction coefficient 430 corresponding to the acquired information about the type 420 of health insurance.

[0020] The second coefficient determination means 200 calculates a second coefficient 450 that increases as the health promotion activities of the single user 400 increase, based on the personal health record 340 of the single user 400. The second coefficient determination means 200 calculates a second coefficient 450 that reflects the health promotion activities of the single user 400, based on the personal health record 340 of the single user 400.

[0021] The third coefficient identification means 210 calculates a third coefficient 460 that increases as the predicted value 360 ​​of medical expenses 350 for one user 400 decreases compared to the average value 370 of medical expenses 350 for the age group to which the one user 400 belongs. The third coefficient identification means 210 calculates a third coefficient 460 that is determined according to the magnitude relationship between the predicted value 360 ​​of medical expenses 350 for one user 400 and the average medical expenses for the age group to which the one user 400 belongs.

[0022] The refund amount calculation means 220 calculates the amount of electronic currency 480 to be refunded to a single user 400 by correcting the standard amount 470 set by the insurer of the health insurance to which the single user 400 belongs based on a first coefficient 440, a second coefficient 450, and a third coefficient 460.

[0023] The refund implementation means 230 performs processing to add the calculated amount of electronic currency 480 to be refunded to one user 400 to the balance of electronic currency 480 managed by one user 400. Based on the above-described operating principle, the device 100 predicts an individual's medical expenses 350 based on information 340 such as daily actions, habits, and test results.

[0024] Furthermore, based on the above-described operating principle, the device 100 returns incentives 480 to individuals in accordance with their behavior in reducing medical expenses, thereby encouraging behavioral changes among the insured 400 and reducing overall medical expenses 350. (Hardware configuration of the information processing device according to this embodiment)

[0025] An example of the hardware configuration of the present device 100 will be described using Fig. 3. Fig. 3 is a diagram showing an example of the hardware configuration of the present device 100. As shown in Fig. 3, the present device 100 has a CPU (Central Processing Unit) 510, a ROM (Read-Only Memory) 520, a RAM (Random Access Memory) 530, an auxiliary storage device 540, a communication I / F 550, an input device 560, an output device 570, and a storage medium I / F 580.

[0026] CPU 510 is a device that executes programs stored in ROM 520, performs arithmetic processing on data loaded into RAM 530 in accordance with program instructions, and controls the entire device 100. ROM 520 stores programs and data to be executed by CPU 510. When CPU 510 executes a program stored in ROM 520, the programs and data to be executed are loaded into RAM 530, and RAM 530 temporarily holds the arithmetic data during the calculation.

[0027] The auxiliary storage device 540 is a device that stores the OS (Operating System), which is basic software, the application program according to this embodiment, and other related data. The auxiliary storage device 540 includes a user identification information storage means 110, a model parameter storage means 120, a correction information storage means 130, and an average medical expense storage means 140, and is, for example, a hard disk drive (HDD) or flash memory.

[0028] The communication I / F 550 is an interface for connecting to a communication network 490 such as a wired or wireless LAN (Local Area Network) or the Internet, and for transmitting and receiving data to and from another device 380 that provides a communication function.

[0029] The input device 560 is a device such as a keyboard for inputting data to the device 100. The output device (display device) 570 is a device formed of an LCD (Liquid Crystal Display) or the like, and functions as a user interface when the user uses the functions of the device 100 or when making various settings. The storage medium I / F 580 is an interface for sending and receiving data to and from a storage medium 590 such as a CD-ROM, DVD-ROM, or USB memory.

[0030] Each of the means included in device 100 may be realized by CPU 510 executing a program corresponding to each of the means stored in ROM 520 or auxiliary storage device 540. Each of the means included in device 100 may also be realized by hardware that performs the processing associated with the means. Alternatively, device 100 may be caused to execute the program by loading the program according to the present invention from an external server device via communication I / F 550 or from storage medium 590 via storage medium I / F 580. (Example of processing by the information processing device according to this embodiment) (1) Medical Cost Prediction Processing by the Device 100

[0031] The flow of medical cost prediction processing by the device 100 will be described with reference to Fig. 4. Fig. 4 is a flowchart showing the flow of an example of medical cost prediction processing by the device 100.

[0032] At S10, the training data collection means 150 acquires from an external device 380 the element data 330 of the training data set 320 to be trained by the machine learning model 310, including the personal health record 340 and medical expenses 350, which are data related to the health and medical care of an individual 400.

[0033] Personal health record 340 includes information on an individual's exercise status and physical condition obtained from wearable devices and mobile information devices worn or operated by the user, health checkup results and medical examination history obtained from the information provision record disclosure system (so-called My Number Portal), an online service operated by the government, and servers managed by health insurance associations and medical institutions. (Past) medical expenses 350 are obtained from the information provision record disclosure system and servers managed by health insurance associations and medical institutions.

[0034] In S20, the training dataset generation means 160 acquires from the user 400 a combination of external device identification information 390 that identifies the external device 380 and user identification information 410 that identifies the individual 400. Then, in S20, the training dataset generation means 160 links the personal health record 340 (330) and medical expenses 350 (330) collected in S10 based on the combination acquired in S20, and integrates them as a training dataset 320.

[0035] In S30, the model generation means 170 trains the machine learning model 310 on the training data set 320 generated in S20, thereby generating the machine learning model 310 that outputs a predicted value 360 ​​of medical expenses 350 when a personal health record 340 is input. The learning algorithm of the machine learning model 310 is not particularly limited.

[0036] In S40, the medical expense prediction means 180 inputs the personal health record 340 for one user 400 into the trained machine learning model 310 generated in S30, and calculates a predicted value 360 ​​of medical expenses 350 for one user 400. By performing the above-described processing, the device 100 predicts the medical expenses 350 of each individual based on information 340 such as daily actions, habits, and test results. (2) Point redemption processing by the device 100

[0037] The flow of the point redemption process by the device 100 will be described with reference to Fig. 5. Fig. 4 is a flowchart showing the flow of an example of the point redemption process by the device 100.

[0038] In S110, the first coefficient identification means 190 acquires information regarding the type 420 of health insurance to which one user 400 is enrolled. Then, in S110, the first coefficient identification means 190 extracts from the correction information storage means 130 the first coefficient 440, which is the correction coefficient 430 corresponding to the information regarding the type 420 of health insurance acquired in S110.

[0039] Furthermore, in S110, the second coefficient determination means 200 calculates a second coefficient 450 that increases as the number of health promotion activities of the single user 400 increases, based on the personal health record 340 of the single user 400. The second coefficient determination means 200 calculates the second coefficient 450 that reflects the health promotion activities of the single user 400, based on the personal health record 340 of the single user 400.

[0040] Furthermore, in S110, the third coefficient identification means 210 calculates a third coefficient 460 that increases as the predicted value 360 ​​of medical expenses 350 for one user 400 calculated in S40 decreases compared to the average value 370 of medical expenses 350 for the age group to which the one user 400 belongs. The third coefficient identification means 210 calculates the third coefficient 460 that is determined according to the magnitude relationship between the predicted value 360 ​​of medical expenses 350 for one user 400 and the average medical expenses for the age group to which the one user 400 belongs.

[0041] At S120, the refund amount calculation means 220 corrects the standard amount 470 set by the insurer of the health insurance to which the user 400 is enrolled based on the first coefficient 440, the second coefficient 450, and the third coefficient 460, and calculates the amount of electronic currency 480 to be refunded to the user 400.

[0042] In S130, the refund implementation means 230 performs a process of adding the calculated amount of electronic currency 480 to be refunded to the one user 400 to the balance of electronic currency 480 managed by the one user 400.

[0043] By performing the above-described processing, the device 100 returns incentives 480 to individuals in accordance with their behavior in reducing medical expenses, thereby encouraging behavioral changes among the policyholders 400 and reducing overall medical expenses 350.

[0044] Although the embodiments of the present invention have been described in detail above, the present invention is not limited to such specific embodiments, and various modifications and variations are possible within the scope of the gist of the present invention as defined in the claims. [Explanation of symbols]

[0045] 100 Information processing device 110 User identification information storage means 120 Model parameter storage means 130 Correction information storage means 140 Mean medical expenses storage means 150 Training data collection methods 160 Training dataset generation method 170 Model Generation Method 180 Medical Cost Prediction Tools 190 First coefficient specifying means 200 second coefficient specifying means 210 Third coefficient specifying means 220 Rebate amount calculation method 230 Means of implementing rebates 310 Machine Learning Models 320 training datasets 330 Element data of the training dataset 340 Personal Health Record 350 Medical expenses 360 Medical Cost Projections 370 Average medical expenses 380 External device 390 External Device Identification Information 400 users 410 User Identity 420 Types of Health Insurance 430 Adjustment coefficients prescribed for each type of health insurance 440 First Coefficient 450 Second Coefficient 460 Third Coefficient 470 Standard amount set by health insurance insurers 480 Electronic currency 490 Communication Networks 510 CPU 520 ROM 530 RAM 540 Auxiliary storage 550 Communication Interface 560 Input Device 570 Output device (display device) 580 Storage Media Interface 590 Storage medium

Claims

1. a learning data collection means for acquiring, from an external device, element data of a learning dataset to be trained by a machine learning model, such as personal health records and medical expenses, which are data related to an individual's health and medical care; a training dataset generation means for acquiring from a user a combination of external device identification information for identifying the external device and user identification information for identifying the individual, and for linking the collected personal health record and medical expenses based on the acquired combination and integrating them as the training dataset; a model generation means for generating the machine learning model that outputs the predicted value of the medical expenses when the personal health record is input by having the machine learning model learn the generated learning dataset; a medical expense prediction means for inputting the personal health record for one of the users into the trained machine learning model and calculating a predicted value of the medical expenses for the one user; a correction information storage means for storing the types of health insurance and the correction coefficients defined for each type of health insurance in association with each other; an average medical expense storage means for storing, for each age group of insured persons under the health insurance, an average medical expense of the insured persons belonging to the age group; a first coefficient specifying means for obtaining information on the type of health insurance to which the one user is enrolled, and extracting, from the correction information storage means, a first coefficient that is the correction coefficient corresponding to the obtained information on the type of health insurance; a second coefficient determining means for calculating a second coefficient that increases as the number of health promotion activities of the user increases based on the personal health record of the user; a third coefficient specifying means for calculating a third coefficient that increases as the predicted value of medical expenses for the one user becomes smaller compared to the average value of medical expenses for the age group to which the one user belongs; a refund amount calculation means for calculating the amount of electronic currency to be refunded to the user by correcting a reference amount set by an insurer of the health insurance to which the user is enrolled based on the first coefficient, the second coefficient, and the third coefficient; and a refund implementation means for performing a process of adding the calculated amount of electronic currency to be refunded to the one user to the balance of electronic currency managed by the one user.

2. a step in which the learning data collection means acquires, from an external device, element data of a learning dataset to be trained by the machine learning model, personal health records and medical expenses, which are data related to an individual's health and medical care; a learning dataset generation means for acquiring from a user a combination of external device identification information for identifying the external device and user identification information for identifying the individual, and linking the collected personal health record and medical expenses based on the acquired combination, and integrating them as the learning dataset; a step in which a model generation means generates a machine learning model that outputs a predicted value of the medical expenses when the personal health record is input by having the machine learning model learn the generated learning dataset; a step in which a medical expense prediction means inputs the personal health record for one of the users into the trained machine learning model and calculates a predicted value of the medical expenses for the one user; An information processing method performed by a computer that includes: a correction information storage means that stores a type of health insurance and a correction coefficient defined for each type of health insurance in association with each other; and an average medical expense storage means that stores, for each age group of insured persons under the health insurance, an average value of medical expenses of the insured persons belonging to the age group; a step in which a first coefficient specifying means acquires information on the type of health insurance to which the one user is enrolled, and extracts a first coefficient, which is the correction coefficient corresponding to the acquired information on the type of health insurance, from the correction information storage means; a step in which a second coefficient specifying means calculates, based on the personal health record of the one user, a second coefficient that increases as the one user engages in more health promotion activities; a step in which a third coefficient specifying means calculates a third coefficient that increases as the predicted value of medical expenses for the one user becomes smaller compared to the average value of medical expenses for the age group to which the one user belongs; a step in which a refund amount calculation means calculates the amount of electronic currency to be refunded to the one user by correcting a reference amount set by an insurer of the health insurance to which the one user is enrolled based on the first coefficient, the second coefficient, and the third coefficient; An information processing method including a step in which a refund implementation means performs a process of adding the calculated amount of electronic currency to be refunded to the one user to the balance of electronic currency managed by the one user.

3. An information processing program for causing a computer to execute the method according to claim 2.

Images