Information processing device, information processing method, and program for predicting fracture risk

Abstract

An information processing device, an information processing method, and a program are provided that provide useful information on fracture risk to users. [Solution] The program causes a computer to function as each means of an information processing device that supports fracture prevention measures, and the information processing device includes a prediction means that predicts the fracture risk of each of multiple subjects based on the subjects' health information, and a selection means that selects subjects for fracture prevention measures based on the fracture risk of the multiple subjects.

Description

[Technical Field]

[0001] The present invention relates to an information processing device, an information processing method, and a program for predicting a fracture risk. [Background technology]

[0002] Among the causes of elderly people needing nursing care, "fractures and falls" are the fourth most common, following "dementia," "cerebrovascular disease (stroke)," and "weakness due to old age" (Ministry of Health, Labor and Welfare, "National Survey of Living Conditions" (2019), URL: https: / / www.mhlw.go.jp / toukei / saikin / hw / k-tyosa / k-tyosa19 / dl / 05.pdf). The leading cause of fractures in the elderly is fragility fractures, and the number of patients with proximal femoral fractures is on the rise, with the five-year mortality rate after proximal femoral fractures exceeding 50% (Osteoporosis Foundation, Osteoporosis in Numbers, URL: https: / / www.jpof.or.jp / osteoporosis / tabid265.html). For this reason, measures to prevent fragility fractures in the elderly are needed (Non-Patent Document 1).

[0003] Patent Document 1 proposes a prediction device that predicts a user's health condition based on feature values ​​including information about fractures, using a prediction model trained using data from the Comprehensive Survey on Living Conditions of the People. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2024-081315 [Non-patent literature]

[0005] [Non-Patent Document 1] Hiroshi Hagino, "The Importance of Preventing Fragility Fractures," Clinical Rheumatology, Japanese Society of Clinical Rheumatology, June 30, 2018, Vol. 30, No. 2, pp. 69-78.<URL=https: / / www.jstage.jst.go.jp / article / cra / 30 / 3 / 30_141 / _pdf / -char / ja> Summary of the Invention [Problem to be solved by the invention]

[0006] The prediction device proposed in Patent Document 1 was unable to predict future fracture risk. Furthermore, although Non-Patent Document 1 mentions the importance of preventing fragility fractures, it does not suggest predicting fracture risk.

[0007] The present invention has been made in view of the above-mentioned problems, and its purpose is to realize a technology that can provide useful information on fracture risk to users. [Means for solving the problem]

[0008] In order to solve this problem, for example, a program according to the present disclosure is a program that causes a computer to function as each means of an information processing device that supports measures against fractures, and the information processing device includes a prediction means for predicting the fracture risk of each of a plurality of subjects based on the health information of the plurality of subjects, and a prediction means for predicting the fracture risk of the plurality of subjects. and one or more of the following conditions: those with a history of fractures who have not been treated for osteoporosis, those who have discontinued treatment for osteoporosis, those who have been judged to require further examination in an osteoporosis screening but have not visited a medical institution, or those who have not undergone osteoporosis screening. and a selection means for selecting individuals to be treated for bone fracture prevention measures based on the results. [Effects of the Invention]

[0009] According to the present invention, it is possible to provide useful information on fracture risk to a user. [Brief explanation of the drawings]

[0010] [Figure 1] 1 is a diagram showing an example of a fracture prevention support system according to an embodiment of the present invention; [Figure 2] FIG. 1 is a block diagram showing an example of the functional arrangement of an information processing apparatus according to an embodiment. [Figure 3] FIG. 2 is a block diagram showing an example of the functional configuration of a terminal device according to the embodiment. [Figure 4] FIG. 2 is a diagram illustrating an example of the functional configuration of a health information database according to the embodiment. [Figure 5] 10 is a flowchart showing a series of processes related to bone fracture prevention according to the embodiment. [Figure 6] 10 is a flowchart showing a process for generating a fracture risk prediction model according to an embodiment. [Figure 7] FIG. 10 is a diagram for explaining the process of generating a fracture risk prediction model according to the embodiment. [Figure 8] FIG. 10 is a diagram illustrating the operation of a fracture risk prediction model according to an embodiment. [Figure 9] 10 is a flowchart showing a process for selecting a target person for fracture prevention according to the embodiment. [Figure 10] FIG. 2 is a diagram illustrating a behavior change probability estimation model according to the embodiment. [Figure 11] 10 is a flowchart showing a process for generating a behavior change probability estimation model according to an embodiment. [Figure 12] 10A and 10B are diagrams illustrating fracture prevention support using behavior change probability according to an embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0011] Hereinafter, the embodiments will be described in detail with reference to the accompanying drawings. Note that the following embodiments do not limit the scope of the invention as claimed, and not all combinations of features described in the embodiments are necessarily essential to the invention. Two or more of the features described in the embodiments may be arbitrarily combined. Furthermore, the same reference numerals are used for the same or similar components, and redundant explanations will be omitted.

[0012] First Embodiment <Configuration of fracture prevention support system> The configuration of a fracture prevention support system according to an embodiment of the present invention will be described with reference to Fig. 1. The fracture prevention support system 100 includes, for example, an information processing device 101, a health information database 102, and a terminal device 103 used by a user. Here, the user is assumed to be an employee of the insurer, but is not limited to this. The information processing device 101, the health information database 102, and the terminal device 103 are communicably connected via a network 104. The network may be any network, such as a LAN, a WAN, or the Internet. All or part of the network may be connected via wireless communication.

[0013] Possible organizations and institutions that implement measures to prevent fracture risk include local governments, insurers (such as national health insurance associations, health insurance societies, and insurers under the Medical Care System for the Elderly, hereinafter referred to as "insurers"), pharmacies, and medical institutions. In this specification, local governments, pharmacies, and medical institutions are collectively referred to as insurers, etc. Insurers, etc. include local governments, health insurance associations, pharmacies, and medical institutions, but may also include other organizations and institutions.

[0014] <Configuration of information processing device> The information processing device 101 is, for example, a server managed by an insurer, etc., or a server managed by a company that provides health data services to insurers, etc. In the following explanation, no distinction is made between insurers, local governments, and service providing companies, and they will be simply referred to as insurers, etc.

[0015] The terminal device 103 is, for example, a user terminal that an insurer or the like uses to access the information processing device 101 when obtaining information related to health guidance. In this embodiment, an example will be described in which the terminal device 103 is, for example, a desktop computer. However, the terminal device 103 is not limited to a desktop computer, and may be any other device that can access the information processing device 101, such as a laptop computer, a tablet terminal, or a smartphone.

[0016] The health information database 102 is a database that stores data related to various types of health information. For example, the health information includes data on health checkups, medical insurance claims, and nursing care certification, but it does not have to include all of these, and may also include other information. The health information is used to train a fracture risk prediction model and to predict fracture risk using the fracture risk prediction model. There may be multiple health information databases 102 for each type of data, or there may be only one. The health information database 102 may be owned by an insurer or other organization. It may be a health insurance claim information / specific health checkup information database (NDB) provided by the Ministry of Health, Labor and Welfare, a database (such as KDB) owned by an examination and payment organization (such as the Social Insurance Medical Fee Payment Fund, the National Health Insurance Association, and the National Health Insurance Federation), or another database owned by a private company. The information processing device 101 may also have a database.

[0017] <Functional configuration of information processing device> Next, an example of the functional configuration of the information processing device 101 will be described with reference to Fig. 2. Note that each of the functional blocks described may be integrated or separated, and the functions described may be realized by different blocks. Also, what is described as hardware may be realized by software, and vice versa.

[0018] The communication unit 201 includes a communication circuit that communicates with various devices via a network. The communication unit 201 receives information processed by the control unit 204 from a communication partner device (e.g., the terminal device 103, etc.), and transmits information processed by the control unit 204 to the communication partner device (e.g., the terminal device 103). The power supply unit 202 is a power supply that provides power required for the operation of the information processing device 101.

[0019] The storage unit 203 includes a non-volatile storage medium such as a hard disk or semiconductor memory, and stores various programs executed by the control unit 204 of the information processing device 101, various data used by the control unit 204, and a DB 230. The various programs include an operating system, frameworks, libraries, and the like, as well as a program for executing fracture risk prediction according to this embodiment. The various data include, for example, setting values ​​of the information processing device 101, data acquired from the health information database 102, data related to the prediction model, learning data, and prediction result data. The storage unit 203 also stores parameters of the prediction model. Data acquired from the health information database 102 may be stored in the DB 230.

[0020] The control unit 204 includes a CPU 210, which is a central processing unit, and a RAM 211. The control unit 204 controls the operation of each unit within the control unit 204 and the operation of each unit of the information processing device 101 by expanding a program stored in the storage unit 203 into the RAM 211 and executing it. The control unit 204 also performs the following functions: predicting fracture risk and selecting recipients for measures to prevent fracture risk.

[0021] The RAM 211 includes a volatile storage medium such as a DRAM, and temporarily stores parameters and processing results for the control unit 204 to execute programs.

[0022] The control unit 204 has functional blocks that are implemented by the CPU 210 reading out a program stored in the storage unit 203 into the RAM 211. The control unit 204 has functional blocks of a prediction model unit 220, a selection unit 221, a learning unit 222, a data acquisition unit 223, a user interface (IF) unit 224, and a notification unit 225.

[0023] The prediction model unit 220 predicts an individual's fracture risk using a fracture risk prediction model. The fracture risk prediction model has a function of inputting data contained in the health information in the health information database 102 and outputting the fracture risk (probability). The prediction period may be the fracture risk within one year, or may be longer or shorter than one year.

[0024] The selection unit 221 selects a subject for guidance and intervention by the insurer or the like based on the individual fracture risk predicted by the prediction model unit 220.

[0025] The learning unit 222 learns the fracture risk prediction model and generates a learned fracture risk prediction model. The learning unit 222 learns the fracture risk prediction model using data acquired from the health information database 102. The learning unit 222 may learn the fracture risk prediction model using, for example, a history of fracture, a diagnosis of osteoporosis, a prescription for osteoporosis medication, a need for further evaluation in an osteoporosis checkup, etc. The learning unit 222 may also use data on fracture hospitalization or consultation due to fracture as correct answer data.

[0026] The data acquisition unit 223 acquires data necessary for predicting fracture risk and selecting targets for measures from the health information database 102 or other data sources. The data acquisition unit 223 may store the acquired data in a database (DB) 230 of the storage unit 203.

[0027] The user IF unit 224 functions as an interface with the terminal device 103 of the insurer, etc. The user IF unit 224 displays an input screen and a setting screen on the terminal device 103, and accepts data input from the terminal device 103.

[0028] The notification unit 225 notifies the selected subject about measures to prevent fractures. The notification unit 225 may notify the subject directly, or may send information about the subject to the terminal device 103. The insurer or the like can notify the selected subject, whose information has been sent to the terminal device 103, about measures to prevent fractures by mail or email.

[0029] <Configuration of terminal device> An example of the functional configuration of the terminal device 103 of the insurer or the like will be described with reference to Fig. 3. Note that each of the functional blocks described may be integrated or separated, and the described functions may be realized by different blocks. Also, what is described as hardware may be realized by software, and vice versa.

[0030] The communication unit 301 includes, for example, a communication circuit, and communicates with the information processing device 101 via mobile communication such as wired LAN or LTE, or via wireless communication such as WiFi, to send and receive the necessary data.

[0031] The operation unit 303 includes buttons and a touch panel provided on the terminal device 103, and accepts operations by a user such as an insurer to display information related to health care services. The display unit 304 includes a display panel such as an LCD or OLED, and displays GUIs for various operations. For example, the display unit 304 displays presentation information generated by the notification unit 225 of the information processing device 101.

[0032] The storage unit 305 includes, for example, a nonvolatile memory such as an HDD or semiconductor memory, and stores programs executed by the control unit 502, etc.

[0033] The control unit 302 includes a CPU 310 and a RAM 311, and controls the operation of each functional block in the control unit 302 and each unit in the terminal device 103 by the CPU 310 executing a program recorded in the storage unit 305, for example.

[0034] <Health information database configuration> Next, an example of the functional configuration of the health information database 102 will be described with reference to FIG. 4. Each of the functional blocks described may be integrated or separated, and the described functions may be implemented by different blocks. Furthermore, what is described as hardware may be implemented by software, and vice versa. The health information database 102 is a database server that stores various health information data according to this embodiment. The health information database 102 includes a communication unit 401, a power supply unit 402, a storage unit 403, and a control unit 404. The communication unit 401 includes, for example, a communication circuit, and communicates with the information processing device 101 via mobile communication such as a wired LAN or LTE, or via wireless communication such as WiFi, to transmit and receive necessary data. The power supply unit 402 is a power source that provides the power required for the operation of the health information database 102.

[0035] The control unit 404 includes a CPU 410 and a RAM 411, and controls the operation of each functional block in the control unit 404 and each unit in the health information database 102, for example, by the CPU 410 executing a program recorded in the storage unit 403.

[0036] The storage unit 403 includes a nonvolatile memory such as an HDD or semiconductor memory, and stores various types of health information data and programs executed by the control unit 502. The storage unit 403 stores health checkup data 420, dental examination (dental checkup) data 430, medical treatment data 440, pharmacy receipt data 450, nursing care certification data 460, intervention data 470, and osteoporosis medical checkup data 480. The health information database 102 does not need to include all of these, and may also include other medical and health-related data. Each piece of data is recorded for each patient (insured person) in association with the date of diagnosis, etc. The medical treatment data 440 and pharmacy receipt data 450 can be acquired from medical fee receipt data. The medical treatment data 440 can be acquired from medical receipt, dental receipt, and nursing care receipt data. Medical receipts, dental receipts, and nursing care receipts are created for each patient, each month of treatment, and by type of inpatient or outpatient visit. The health information data 102 may store health checkup data 420, dental examination (dental checkup) data 430, medical treatment data 440, prescription receipt data 450, nursing care certification data 460, intervention data 470, and osteoporosis screening data 480, linked to a ledger. In recent years, with the use of the My Number card as a health insurance card, prescription data has been recorded in association with the My Number (individual number). Therefore, the My Number allows for the acquisition of comprehensive data related to personal medical information, such as records of health checkup data 420, dental examination data 430, medical treatment data 440, prescription receipt data 450, nursing care certification data 460, and osteoporosis screening data 480, as well as vital signs data acquired from wearable devices and electronic medical records.

[0037] The health information database 102 may also acquire and use data from the Ministry of Health, Labor and Welfare's Health Insurance Claim Information and Specific Health Checkup Information Database (NDB), databases (such as the KDB) owned by screening and payment agencies (such as the Social Insurance Medical Fee Payment Fund, the National Health Insurance Association, and the National Health Insurance Federation), and other databases owned by private companies. The health information database 102 may also be the NDB, the KDB, or other databases owned by private companies. In the NDB, patient (insured) information is anonymized, but different types of patients are linked by the same hash ID. The data acquisition unit 223 of the information processing device 101 may also be configured to access the NDB to acquire data. The learning unit 222 and the selection unit 221 of the information processing device 101 can use data from the NDB, the KDB, and other databases.

[0038] Each data item will be explained below. The health checkup data 420 is data on health checkups for a certain population. Municipalities and corporate health insurance associations conduct health checkups for their residents and members. The population of the health checkup data 420 may be a single municipality or a single company. The population of the health checkup data 420 may also be multiple municipalities or multiple companies. In the following, an example will be described in which the insurer is a company that provides health guidance services and uses the health checkup data 420 from multiple municipalities, multiple health insurance associations, etc. after obtaining permission to use the data. The same applies to the populations of the dental checkup data 430, medical treatment data 440, prescription data 450, nursing care certification data 460, intervention data 470, and osteoporosis checkup data 480 described below.

[0039] The health checkup data 420 is data on health checkups that a subject has undergone, associated with the subject and accumulated by year. The health checkup data 420 stores information on whether the subject has undergone a health checkup, as well as measurement values ​​for each health checkup item. The health checkup data 420 includes interview (questionnaire) data that the subject answers in advance before the health checkup. The health checkup data includes specific health checkup data, which is data on the results of health checkups conducted for subjects aged 40 to 74 with a focus on metabolic syndrome, and late-stage elderly health checkup data conducted for subjects aged 75 or older. The health checkup data 420 may also include health checkup data conducted by insurers and other organizations for their employees in accordance with the Industrial Safety and Health Act. The health checkup data 420 includes data such as examinee information, specific health checkup result information, interview (questionnaire) information (medication, smoking history, etc.), whether the subject meets the metabolic syndrome criteria, and whether the subject meets the eligibility criteria for specific health guidance. Of the 420 health checkup data of the subjects, specific health checkup test items include health questions (medication history, smoking history), questionnaires, anthropometric measurements (height, weight, BMI, waist circumference), physical examination (physical examination), urinalysis (urine glucose, urinary protein), blood tests (lipids (triglycerides, HDL cholesterol, LDL cholesterol), glucose metabolism (fasting blood glucose or hemoglobin A1C), and liver function (GOT, GPT, γ-GTP)), and additional items include anemia tests (red blood cell count, hemoglobin content, hematocrit value), electrocardiogram, fundus examination, and renal and urinary tract tests.

[0040] The dental examination data 430 is data on dental examinations accumulated by year in association with the subject. The dental examination data 430 may also include data on dental examinations such as regular health checkups and specific health checkups. The dental examination data 430 accumulates whether the subject has undergone a dental examination or not, as well as the results of each dental examination item. The dental examination data 430 may also include questionnaire data that the subject answers when undergoing a dental examination.

[0041] The medical data 440 is data that records the subject's visits to medical institutions and their admissions and discharges. It can be obtained from medical and dental receipt data created for each inpatient and outpatient visit. The medical data 440 includes the name of the medical institution, medical history, date of treatment, and name of medical procedure.

[0042] The dispensing receipt data 450 includes information on drugs dispensed at a dispensing pharmacy, information on the prescribing source, and information on the patient. The dispensing receipt data 450 is associated with each patient. The dispensing receipt data 450 includes the name of the medical institution or pharmacy, the date of dispensing, the name of the drug, the names of ingredients, usage, dosage, etc.

[0043] The nursing care certification data 460 is data of information related to nursing care certification. Nursing care certification is determined by a nursing care certification review board attached to a local government, based on an application for review, by calculating the standard time for nursing care certification etc. for five areas (direct assistance for daily living, indirect assistance for daily living, BPSD-related activities, functional training-related activities, and medical-related activities) and determining whether a person needs support level 1 to level 5 based on the total of the calculated time and the dementia surcharge. The nursing care certification data 460 may also include "independent," which is a state in which support such as nursing care services is not required for daily living. The nursing care certification data 460 may also include support level data.

[0044] Intervention data 470 includes data recording measures implemented by insurers, etc., correlating the implementer, implementation date, and details, personal medical data obtainable from electronic medical records, and vital data obtainable from wearable devices, etc. Interventions are divided into individual and group interventions. An example of individual intervention is specific health guidance for those with metabolic syndrome or those at risk of metabolic syndrome. Examples of group interventions include encouraging those who have not yet undergone health checkups to do so, and health promotion efforts by insurers, etc. Regarding osteoporosis, examples include encouraging people to undergo osteoporosis screening and encouraging those who require further evaluation to visit a medical institution. This fracture prevention support system handles both group and individual interventions.

[0045] Osteoporosis Screening Data 480 records osteoporosis screening data. Under the Health Promotion Act, osteoporosis screening is conducted for women aged 40 to 70 in five-year increments. Some local governments independently expand the scope of screening to include a wider range of genders and ages. Osteoporosis screening is also conducted by health insurance associations and other organizations. Osteoporosis screening may include a medical interview, bone density testing, physical measurements, and blood and urine tests. The medical interview involves filling out information about lifestyle and dietary habits, allowing researchers to understand the subject's nutritional and exercise status. Bone density tests are available using DXA, MD, and ultrasound, depending on the location and whether or not they involve radiation. Physical measurements involve continuously measuring the subject's height; any loss of height indicates a suspicion of osteoporosis. Blood and urine tests measure bone metabolic markers, which provide an indicator of bone metabolism.

[0046] <Machine learning model> The fracture risk prediction model for predicting fracture risk in this embodiment may use a machine learning model. Various types of data in the health information database 102 are used as explanatory variables. The data used in the health information database 102 may be some types of data or all types of data. The objective variable of the machine learning model is fracture risk.

[0047] <Overall flow of this system> The overall flow of this embodiment will be described with reference to Fig. 5. The flowchart in Fig. 5 is implemented by the CPU 210 of the information processing device 101 loading a program stored in the storage unit 203 into the RAM 211 and executing it. In the following, the step numbers of each process included in the flowchart are indicated by numbers beginning with "S". This also applies to the subsequent flowcharts.

[0048] In S501, the learning unit 222 of the information processing device 101 generates a fracture risk prediction model. The fracture risk prediction model is a model that receives input of medical examination data and the like for an individual subject and outputs a fracture risk as a probability. The fracture risk prediction model may also output a fracture risk, including secondary fractures and domino fractures, as described below, as a probability. Details of S501 will be described later.

[0049] Next, in S502, the prediction model unit 220 of the information processing device 101 performs a process of using the prediction model to obtain the probability of fracture risk for each subject in a certain population. Then, based on the individual fracture risk predicted by the prediction model unit 220, the selection unit 221 selects subjects for whom measures against the fracture risk should be taken. For example, subjects for whom measures should be taken are selected in descending order of fracture risk. Details of S502 will be described later.

[0050] Finally, in S503, the notification unit 225 of the information processing device 101 notifies the high-risk subjects selected by the selection unit 221. Then, the user IF unit 224 transmits information about the selected subjects to the terminal device 103. The terminal device 103 can display information about subjects with a high fracture risk on the display unit 304. By examining the information about subjects who are candidates for measures against fracture risk, insurers and the like can determine whether measures against fracture risk are necessary for the subjects. The content of the notification by the notification unit 225 of the information processing device 101 may include not only information about subjects with a high fracture risk but also effective measures for the subjects. For example, subjects who have discontinued treatment or outpatient visits may be encouraged to visit a medical institution or the like again. Furthermore, if the risk of fracture is high, possible measures include a recommendation for a bone density test or other examination, starting osteoporosis medication and providing instructions on taking the medication, recommending exercise, and improving nutrition. Details of S503 will be described later.

[0051] <Fragility fracture> Before explaining the fracture risk prediction model, we will explain fragility fractures and their risks. Fractures include traumatic fractures, which occur when a strong external force is temporarily applied during a traffic accident, sports, or other such event, and fragility fractures, which occur when bone strength is weakened due to osteoporosis or other factors, leading to bone breakage due to a small external force, such as a fall from standing height. The fractures targeted by the fracture prevention support system of this embodiment are fragility fractures. The small external force referred to in fragility fractures generally refers to a force weaker than a fall from standing height. For example, this includes a weak force such as falling on one's hands or buttocks, or lifting a heavy object. While such cases would not normally result in a fracture, a weakened bone strength can lead to a fracture.

[0052] Examples of fragility fractures include distal radius fractures (wrist fractures), proximal humerus fractures (fractures at the base of the shoulder), thoracolumbar vertebral fractures (compression fractures of the spine), and proximal femur fractures (fractures at the base of the leg).Proximal femur fractures and thoracolumbar vertebral fractures in particular can lead to bedriddenness and affect life prognosis, so prevention and measures are important.

[0053] Fragility fractures are caused by the fragility of bones throughout the body, which can lead to further fractures and secondary fractures. Fragility fractures occur repeatedly, sometimes referred to as domino fractures. Secondary and domino fractures are thought to occur because, once a fracture occurs, the body cannot move while the fracture is healing, leading to a vicious cycle of repeated fractures due to a decrease in muscle strength and a decrease in stimulation to the bone, which further weakens the bones.

[0054] For fragility fractures, subjects can be extracted from the medical data 440 based on a history of distal radius fractures, proximal humerus fractures, thoracic and lumbar vertebral fractures, and proximal femur fractures. In addition, for osteoporosis, which is a cause of fragility fractures, subjects can be extracted based on the diagnosis name in the medical data 440 and the osteoporosis treatment medication in the prescription data 450. In addition, subjects can be extracted from the osteoporosis screening data 480 based on whether they have been determined to require detailed examination, or if they are people who are prioritizing measures for osteoporosis screening, such as people with diabetes.

[0055] Osteoporosis has no noticeable symptoms, so it is often discovered late, and even if treatment is started, it is often discontinued midway. When fragility fractures occur, patients often become bedridden and require nursing care, which causes medical expenses and nursing care benefits to rise.

[0056] <Creating a fracture risk prediction model> Details of S501 in FIG. 5 will be described with reference to FIGS. 6 to 8. First, generation of a fracture risk prediction model will be described with reference to the flowchart in FIG. 6 and the relationship between the data and the model in FIG. 7. The flowchart in FIG. 6 is implemented by the CPU 210 of the information processing device 101 loading a program stored in the storage unit 203 into the RAM 211 and executing it. Here, the health check data 420, dental check data 430, medical treatment data 440, prescription data 450, and osteoporosis check data 480 in the health information database 102 are used as training data, but some of these may be used, or other data may be included. Data may be appropriately selected from the health information database 102 and used as training data.

[0057] In S601, the data acquisition unit 223 of the information processing device 101 acquires the health checkup data 420, dental checkup data 430, medical treatment data 440, prescription data 450, and osteoporosis checkup data 480 for fiscal year 2021 from the health information database 102 and stores them in the memory unit 203.

[0058] Next, in S602, the data acquisition unit 223 of the information processing device 101 acquires the medical data 440 for fiscal year 2022 from the health information database 102 and stores it in the memory unit 203. In S602, the learning unit 222 may acquire patterns of health checkup values, questionnaires, medical history, prescription drugs, and osteoporosis checkups for fiscal year 2022.

[0059] Next, in S603, the learning unit 222 identifies subjects who were hospitalized or examined for a fracture in fiscal year 2022. The learning unit 222 may also identify subjects who were hospitalized or examined for a fracture in fiscal year 2021, in addition to fiscal year 2022.

[0060] Next, in S604, the learning unit 222 trains the fracture risk prediction model with the patterns of medical checkup values, questionnaires, medical history, prescribed medications, and osteoporosis screening data from fiscal year 2021 for subjects who were hospitalized or examined for a fracture in fiscal year 2022. The learning unit 222 may extract subjects from the medical data 440 as subjects who were hospitalized or examined for a fracture based on a history of distal radius fracture, proximal humerus fracture, thoracic and lumbar vertebral fracture, or proximal femur fracture. The learning unit 222 may extract subjects from the medical data 440 as subjects who were hospitalized or examined for a fracture based on the diagnosis name in the medical data 440 and the osteoporosis medication in the prescription data 450. The learning unit 222 may also extract subjects who were determined to require further examination in the osteoporosis screening data 480 as subjects who were hospitalized or examined for a fracture.

[0061] This allows the fracture risk prediction model to learn what characteristics subjects have that make them at high risk of fracture. In S604, the learning unit 222 may train the fracture risk prediction model on the patterns of health checkup values, questionnaires, medical history, prescribed medications, and osteoporosis checkups in fiscal year 2021 for subjects who were hospitalized or examined for fractures in fiscal year 2021. The learning algorithm uses the well-known gradient boosting, but is not limited to this.

[0062] In S602, the learning unit 222 may acquire patterns of health checkup values, questionnaires, medical history, prescription medications, and osteoporosis checkups for fiscal year 2022. Then, the learning unit 222 may omit S601 and train a fracture risk prediction model using patterns of health checkup values, questionnaires, medical history, prescription medications, and osteoporosis checkups for fiscal year 2022 in S604. Furthermore, in S604, the learning unit 222 may train a fracture risk prediction model using patterns of health checkup values, questionnaires, medical history, prescription medications, and osteoporosis checkups for both fiscal years 2021 and 2022.

[0063] As a result of the learning performed by the learning unit 222 in S604, a fracture risk prediction model is generated in S605.

[0064] In training the fracture risk prediction model described above, health data from fiscal years 2021 and 2022 was used, but health data from any fiscal year can be used. Furthermore, the period does not have to be a fiscal year, and the period does not have to be consecutive. The period can be a calendar year or a period of two years or more. It can also be in monthly or several-month increments.

[0065] The medical interview (questionnaire) data included in the health checkup data 420 includes information on the subject's responses to questions such as whether or not they use medication, including blood pressure lowering medication, insulin injections or blood sugar lowering medication, cholesterol lowering medication, medical history such as stroke, heart disease, and kidney failure, smoking habits, weight gain since age 20, exercise habits, eating habits, drinking habits, and sleep status. Because this information in the medical interview (questionnaire) data is self-acknowledged by the subject, it serves as an index for risk prediction using a fracture risk prediction model. For example, people who do not exercise regularly may have low bone density and tend to have a higher fracture risk. Furthermore, drinking habits increase the risk of falls due to alcohol consumption, so they also tend to have a higher fracture risk.

[0066] FIG. 8 shows the operation of the fracture risk prediction model generated by the learning unit 222. The fracture risk prediction model generated by the flow of FIG. 6 outputs a probability of fracture risk for a certain subject when the subject's health checkup data, dental examination data, medical data, prescription data, and osteoporosis screening data are input. The input and output of the fracture risk prediction model do not necessarily match the training data used to train the fracture risk prediction model. For example, the fracture risk prediction model may be trained using five years' worth of data, and the subject's health information may be input and the fracture risk may be output. The prediction period may be the fracture risk within one year, or may be shorter or longer than one year. Furthermore, the fracture risk may be predicted by inputting health information for multiple years. The fracture risk may also be predicted by inputting health information for multiple discontinuous periods.

[0067] <Target for fracture prevention> The detailed processing of S502 in Fig. 5 will be described with reference to Fig. 9. The flowchart in Fig. 9 is implemented by the CPU 210 of the information processing device 101 reading a program stored in the storage unit 203 into the RAM 211 and executing it.

[0068] The fracture risk prediction model outputs an individual's risk of fracture as a probability. By using the fracture risk prediction model, it is possible to predict a subject's fracture risk (probability) by inputting their health checkup data, dental checkup data, medical data, pharmacy prescription data, and osteoporosis screening data. The prediction period may be the fracture risk within one year, or it may be shorter or longer than one year. The fracture risk prediction model in this embodiment can predict not only the fracture risk of subjects with no history of fracture, but also the risk of recurrent fractures, secondary fractures, and domino fractures.

[0069] Here, the population for which fracture risk is predicted does not necessarily have to be the same as the population of subjects used to generate the fracture risk prediction model. Since accuracy is generally required for generating a prediction model, the more data there is, the better. On the other hand, risk prediction using a prediction model requires the selection of subjects taking into account factors such as the size of the insurer and the composition of the subjects, so this is the population for which the insurer will implement measures against fractures.

[0070] In S901, the prediction model unit 220 of the information processing device 101 first predicts the fracture risk (probability) of each individual for all subjects of the insurer's or other person's fracture prevention measures using the fracture risk prediction model.

[0071] Next, in S902, the selection unit 221 of the information processing device 101 ranks the subjects in descending order of risk based on their fracture risk. Then, in S903, the selection unit 221 selects subjects with a high risk. The selection unit 221 may select subjects with a predetermined risk or higher, may select subjects until a predetermined number is reached, or may select subjects at a predetermined ratio to all subjects.

[0072] In addition to the fracture risk (probability of fracture), the ranking in S902 may be based on the following ranking from the perspective of encouraging a patient to visit a medical institution and leading to treatment. In this case, the risk prediction using the fracture risk prediction model in S901 may be omitted. (1) From the medical data 440, select untreated individuals with a history of a distal radius fracture, a proximal humerus fracture, a thoracic / lumbar vertebral fracture, or a proximal femur fracture. (2) Select subjects who have discontinued treatment for osteoporosis. (3) Select subjects who have been determined to require further examination for osteoporosis but have not yet visited a medical institution. (4) Select subjects who are at high risk of fracture but have not undergone osteoporosis screening.

[0073] The learning unit 222 can also train the fracture risk prediction model to increase the risk of subjects in the above categories (1) to (4). The selection unit 221 may prioritize subjects in the above categories (1) to (4) as targets for fracture prevention measures, and output the targeted factors as reference information for the selected subjects. By referring to the reference information for the selected subjects, insurers and the like can implement effective fracture prevention measures for the subjects.

[0074] <Second embodiment> In the first embodiment, subjects with a high risk of fracture are selected as subjects for whom measures to prevent fractures are to be taken. In the second embodiment, subjects for whom measures to prevent fractures are to be taken are selected by using the probability of behavioral change in addition to the level of fracture risk.

[0075] Behavioral change refers to a change in a subject's behavior as a result of intervention by an insurer, etc. An example would be a subject who has not visited a medical institution, etc., receiving a recommendation from an insurer, etc. to do so and beginning to do so. Another example would be a subject receiving guidance from an insurer, etc. to improve their lifestyle habits and working to improve their lifestyle habits. The probability of behavioral change is the probability that a subject's behavior will change as a result of intervention by an insurer, etc. An example would be the probability that a subject will visit a medical institution, etc., as a result of an insurer, etc. recommending the subject to visit a medical institution, etc.

[0076] When an insurer etc. encourages a target individual to visit a medical institution etc., the target individual may respond differently, such as completely ignoring the recommendation or not seeing treatment or improvement even if they do. It also depends on the content of the recommendation made by the insurer etc. The way the target individual perceives the recommendation and their subsequent behavior will be different depending on whether the insurer etc. sends direct mail to the target individual or whether the insurer etc. makes a phone call or visits the individual to encourage them to visit a medical institution etc. Furthermore, people who have never had a health checkup and people who receive a health checkup every year have different awareness of health and will behave differently in response to the recommendation.

[0077] <Generation of a behavior change probability estimation model> In this embodiment, in response to intervention by an insurer or the like, the health checkup data 420 and medical care data 440 of a subject who has visited a medical institution or the like are used as training data to train a behavior change probability estimation model.

[0078] The behavior change probability estimation models of this embodiment include a consultation probability model, a response model, and an uplift model. Each of these models will be described with reference to FIG.

[0079] FIG. 10(a) is a conceptual diagram illustrating the generation of a medical examination probability model. In the medical examination probability model, health checkup data 420, dental examination data 430, medical treatment data 440, pharmacy receipt data 450, and osteoporosis examination data 480 are used as learning data. The data acquisition unit 223 of the information processing device 101 acquires the health checkup data 420, dental examination data 430, medical treatment data 440, pharmacy receipt data 450, and osteoporosis examination data 480 from the health information database 102 and stores them in the storage unit 203. The health checkup data 420, dental examination data 430, medical treatment data 440, pharmacy receipt data 450, and osteoporosis examination data 480 acquired by the data acquisition unit 223 may be from any period. Furthermore, only a portion of these data may be used, or other data may be included.

[0080] Next, the learning unit 222 identifies subjects who have actually taken action. For example, subjects who have undergone a health checkup are identified. The learning unit 222 identifies attributes that tend to lead to undergoing a health checkup from the health checkup data 420, dental checkup data 430, medical care data 440, and prescription data 450 of the subjects who have undergone the health checkup, and generates a checkup probability model. Instead of or in addition to a health checkup, subjects who have undergone a dental checkup may be identified.

[0081] As a result, when a subject's health checkup data 420, dental checkup data 430, medical treatment data 440, and prescription data 450 are input into the medical examination probability model, a probability indicating a tendency to undergo a health checkup is output. Here, it is assumed that subjects who undergo a health checkup also have a high tendency to visit a medical institution, etc., and the output of the medical examination probability model is the probability that a subject will visit a medical institution, etc. The medical examination probability model is an estimation model that outputs the probability that a subject will visit a medical institution, etc., even if an insurer, etc., does not take any intervention to recommend the subject.

[0082] Next, the response model will be described. FIG. 10(b) is a conceptual diagram illustrating the generation of a response model. In the response model, health checkup data 420, dental examination data 430, medical treatment data 440, pharmacy receipt data 450, osteoporosis examination data 480, and intervention data 470 are used as learning data. The data acquisition unit 223 of the information processing device 101 acquires the health checkup data 420, dental examination data 430, medical treatment data 440, pharmacy receipt data 450, osteoporosis examination data 480, and intervention data 470 from the health information database 102 and stores them in the storage unit 203. The health checkup data 420, medical treatment data 440, pharmacy receipt data 450, and intervention data 470 acquired by the data acquisition unit 223 may be from any period. Furthermore, only a portion of these data may be acquired, or other data may be included.

[0083] The intervention data 470 includes, for example, data in which an insurer or the like notifies a subject of a recommendation to visit a medical institution or the like based on the subject's health checkup data 420, etc. The learning unit 222 identifies a subject whose behavior has changed due to an intervention based on the intervention data 470. Here, as an example, an insurer or the like recommends a subject to visit a medical institution or the like, and identifies a subject who visited the medical institution or the like in accordance with the recommendation. The learning unit 222 identifies attributes that indicate a tendency to visit a medical institution or the like from the health checkup data 420, dental examination data 430, medical care data 440, prescription data 450, and osteoporosis health checkup data 480 of the subject who visited the medical institution or the like in accordance with the recommendation, and generates a response model. In this case, the learning unit 222 may use the health checkup data 420, dental checkup data 430, medical care data 440, prescription data 450, and osteoporosis checkup data 480 of subjects who were encouraged by an insurer to visit a medical institution, etc. but did not visit a medical institution, etc., in addition to the attributes that indicate a tendency to visit a medical institution, etc., to identify attributes that make them less likely to visit a medical institution, etc., and use these as learning data to generate a response model.

[0084] When a subject's health checkup data 420, dental checkup data 430, medical treatment data 440, pharmacy prescription data 450, and osteoporosis checkup data 480 are input into the response model, the output is the probability that the subject will actually visit a medical institution, etc. if the subject is encouraged to visit a medical institution, etc. The response model has the advantage of being able to generate an estimation model even if the insurer, etc. has a low track record of encouraging patients to visit a medical institution, etc.

[0085] Next, the uplift model will be described with reference to Fig. 10(c) and Fig. 11. Fig. 10(c) is a conceptual diagram for explaining generation of the uplift model. The flowchart in Fig. 11 is realized by the CPU 210 of the information processing device 101 reading a program stored in the storage unit 203 into the RAM 211 and executing it.

[0086] In the uplift model, the health check data 420, dental examination data 430, medical treatment data 440, pharmacy receipt data 450, osteoporosis examination data 480, and intervention data 470 are used as learning data. In S1101, the data acquisition unit 223 of the information processing device 101 acquires the health check data 420, dental examination data 430, medical treatment data 440, pharmacy receipt data 450, and osteoporosis examination data 480 from the health information database 102 and stores them in the storage unit 203. The health check data 420, dental examination data 430, medical treatment data 440, pharmacy receipt data 450, and osteoporosis examination data 480 acquired by the data acquisition unit 223 may be from any period.

[0087] In S1102, the data acquisition unit 223 of the information processing device 101 acquires intervention data 470 and stores it in the storage unit 203. The intervention data 470 includes, for example, data in which an insurer or the like notifies a subject of a recommendation to visit a medical institution or the like based on the subject's health checkup data 420 or the like. Here, the intervention data 470 includes, for example, data on subjects who were not recommended to visit a medical institution or the like, data on subjects who were recommended to visit a medical institution or the like by a regular paper notification, and data on subjects who were recommended to visit a medical institution or the like by a detailed paper notification. The recommendation notification is not limited to these, and may include, for example, a telephone call, an email, or a visit, or may be a notification with further changes to the content of the paper notification. The intervention data 470 prepared as learning data may be a notification indicating whether or not a recommendation to visit a medical institution or the like was made, or a notification indicating a recommendation at multiple different levels.

[0088] Next, in S1103, the learning unit 222 of the information processing device 101 causes the uplift model to learn the attributes of subjects who actually visited a medical institution, etc., for subjects who were not notified (non-intervention) of a recommendation to visit a medical institution, etc. This enables the uplift model to estimate the probability of visiting a medical institution, etc., according to the attributes of the subject in the non-intervention case.

[0089] Next, in S1104, the learning unit 222 of the information processing device 101 causes the uplift model to learn the attributes of subjects who have actually visited a medical institution, etc., after being encouraged to visit a medical institution, etc. through a written notification (intervention). This allows the uplift model to estimate the probability of visiting a medical institution, etc., depending on the attributes of the subject, when intervention is performed.

[0090] Next, in S1105, the learning unit 222 of the information processing device 101 compares the consultation probability in the case of no intervention in S1103 with the consultation probability in the case of a paper notification (intervention) in S1104. For example, if the consultation probability estimated in S1104 is higher than the consultation probability estimated in S1103, the learning unit 222 can determine that the paper notification (intervention) is effective, excluding the influence of non-intervention. Since highly health-conscious subjects will naturally visit a medical institution or the like even without intervention from an insurer or the like, in S1105, the effect of the intervention, excluding the influence of no intervention, and the increase in the probability of visiting a medical institution or the like due to the intervention can be calculated.

[0091] In S1106, the learning unit 222 trains the uplift model to learn the difference between the consultation probability in the case of no intervention in S1105 and the consultation probability in the case of intervention, according to the attributes of the subject. In the example described above, the learning unit 222 compared the consultation probability in the case of no intervention with the consultation probability in the case of intervention, but it may also compare the consultation probability in the case of a normal notification with the consultation probability in the case of a detailed notification. In this case, it is possible to determine the extent to which the probability of not visiting a medical institution, etc. with a normal notification but visiting a medical institution, etc. increases with a detailed notification.

[0092] When a subject's health checkup data 420, dental checkup data 430, medical treatment data 440, pharmacy prescription data 450, and osteoporosis checkup data 480 are input into the generated uplift model, the increase in the probability that the subject will actually visit a medical institution when the subject is encouraged to visit a medical institution is output.

[0093] The learning unit 222 may further verify and re-learn the generated uplift model. In S1107, the learning unit 222 inputs the health check data 420, the medical care data 440, and the prescription data 450, and outputs the increase in the probability that the subject will visit a medical institution or the like. Then, the learning unit 222 may re-learn the uplift model depending on whether the subject who has received the intervention actually visits (takes action at) a medical institution or the like. If the subject actually visits (takes action at) a medical institution or the like, the uplift model may be corrected to increase its output probability, and if the subject does not visit (takes action at) a medical institution or the like for a certain period of time, the uplift model may be corrected to decrease its output probability.

[0094] According to the behavior change probability estimation model, by inputting health information, it is possible to output the probability of a subject's behavior change. For example, the behavior change probability estimation model can estimate the following: Older subjects who underwent a health checkup in the previous year are more likely to undergo a health checkup in response to a recommendation. Subjects who have not undergone a health checkup in the past 1-2 years, or subjects who only undergo a health checkup once every few years, are more likely to undergo a health checkup when recommended to do so. Furthermore, subjects who have never undergone a health checkup in the past are less likely to undergo a health checkup even when recommended to do so. Subjects who visit medical institutions with a certain frequency are more likely to respond to recommendations. Elderly people are more likely to respond to recommendations to undergo a health checkup via telephone. In this way, the behavior change probability estimation model outputs the probability of behavior change by inputting health information data.

[0095] <Selection of subjects using a behavior change probability estimation model> Next, the selection of subjects for which measures are to be taken using a fracture risk prediction model and a behavioral change probability estimation model will be described with reference to Figure 12. The behavioral change probability estimation model may be any of a consultation probability model, a response model, and an uplift model. In the second embodiment, subjects are selected by combining the behavioral change probability estimation model with the fracture risk prediction model of the first embodiment.

[0096] First, the prediction model unit 220 of the information processing device 101 inputs the subject's individual medical checkup data, dental checkup data, medical treatment data, prescription data, and osteoporosis checkup data (1201) into a fracture risk prediction model 1202. Then, a fracture risk (probability) 1203 for a certain subject is output. The prediction period may be the fracture risk for one year (or less), or may be longer or shorter than one year.

[0097] Furthermore, the prediction model unit 220 of the information processing device 101 inputs the subject's individual health checkup data, dental checkup data, medical treatment data, prescription data, and osteoporosis checkup data (1204) into a behavior change probability estimation model 1205. This then obtains, for a certain subject, a probability 1206 of visiting a medical institution or the like when an insurer or the like intervenes.

[0098] Next, the prediction model unit 220 calculates the probability 1207 of a subject visiting a medical institution or the like as a result of an intervention, by multiplying the fracture risk (probability) 1203 of a certain subject by the probability 1206 of the subject visiting a medical institution or the like as a result of an intervention. The calculation performed by the prediction model unit 220 may also be to weight either the fracture risk 1203 or the behavior change probability 1206 and then take the product.

[0099] By reflecting the probability of behavioral change due to intervention in the fracture risk predicted by the fracture risk prediction model, it is possible to identify subjects who are at high risk of fracture and in need of countermeasures, and who are likely to visit a medical institution or the like if an insurer or the like intervenes. After determining the probability 1207 of visiting a medical institution or the like as a result of intervention for the fracture risk of each subject, subjects for whom countermeasures will be taken are selected according to the flow in Figure 9. According to the second embodiment, it is possible to select subjects for whom countermeasures will be more effective.

[0100] The invention is not limited to the above-described embodiment, and various modifications and variations are possible within the scope of the gist of the invention. [Explanation of symbols]

[0101] 101: Information processing device, 102: Health information database, 103: Terminal device

Claims

1. A program that causes a computer to function as each means of an information processing device that supports bone fracture prevention, the information processing device comprising: A prediction means for predicting a fracture risk of each of a plurality of subjects based on the health information of the plurality of subjects; a selection means for selecting subjects for fracture prevention measures based on the fracture risk of the plurality of subjects and one or more of the following conditions: subjects with a history of fracture who have not yet been treated for osteoporosis, subjects who have discontinued treatment for osteoporosis, subjects who have been determined to require further examination in an osteoporosis screening but have not yet visited a medical institution, or subjects who have not undergone an osteoporosis screening; A program that includes:

2. A program that causes a computer to function as each means of an information processing device that supports bone fracture prevention, the information processing device comprising: A prediction means for predicting a fracture risk of each of a plurality of subjects based on the health information of the plurality of subjects; an estimation means for estimating a probability of behavioral change due to an intervention for each of the plurality of subjects based on health information of the plurality of subjects; a selection means for selecting subjects for bone fracture prevention measures based on the bone fracture risk and the behavioral change probability of the plurality of subjects; A program that includes:

3. The prediction means predicts the fracture risk using a prediction model trained based on past health information of the subject who has had a fracture. The program according to claim 1 or 2.

4. the selection means selects people to be targeted for countermeasures until the number of people selected exceeds a predetermined number or a predetermined ratio. The program according to claim 1 or 2.

5. The selection means outputs the selected subject together with the factors that led to the selection of the selected subject. The program according to claim 1 or 2.

6. The estimation means includes an estimation model that learns based on intervention information and behavioral change information. The program according to claim 2.

7. The program according to claim 6, wherein the estimation model divides subjects into two or more groups depending on whether or not an intervention is performed or the level of the intervention, and learns based on whether or not there is a behavioral change in the subjects in each group.

8. An information processing device that supports bone fracture prevention, A prediction means for predicting a fracture risk of each of a plurality of subjects based on the health information of the plurality of subjects; and a selection means for selecting subjects for fracture prevention measures based on the fracture risk of the plurality of subjects and one or more of the following conditions: subjects with a history of fracture who have not yet been treated for osteoporosis, subjects who have discontinued treatment for osteoporosis, subjects who have been determined to require further examination in an osteoporosis screening but have not yet visited a medical institution, or subjects who have not undergone osteoporosis screening. Information processing device.

9. An information processing device that supports bone fracture prevention, A prediction means for predicting a fracture risk of each of a plurality of subjects based on the health information of the plurality of subjects; an estimation means for estimating a probability of behavioral change due to an intervention for each of the plurality of subjects based on health information of the plurality of subjects; and a selection means for selecting subjects for bone fracture prevention measures based on the bone fracture risk and the behavioral change probability of the plurality of subjects. Information processing device.

10. An information processing method executed by an information processing device that supports bone fracture prevention, a prediction step of predicting a fracture risk for each of a plurality of subjects based on health information of the plurality of subjects; and a selection step of selecting subjects for fracture prevention measures based on the fracture risk of the plurality of subjects and one or more conditions of untreated osteoporosis among those with a history of fracture, those who have discontinued treatment for osteoporosis, those who have been determined to require further examination in an osteoporosis screening but have not visited a medical institution, or those who have not undergone osteoporosis screening. Information processing methods.

11. An information processing method executed by an information processing device that supports bone fracture prevention, a prediction step of predicting a fracture risk for each of a plurality of subjects based on health information of the plurality of subjects; an estimation step of estimating a probability of behavior change due to an intervention for each of the plurality of subjects based on health information of the plurality of subjects; and a selection step of selecting subjects for a bone fracture prevention measure based on the bone fracture risk and the behavior change probability of the plurality of subjects. Information processing methods.

Images