Service providing method regulating system and method

Abstract

To provide a convenient service providing method regulating system and method.SOLUTION: A service providing method regulating system 1000 regulating a method for providing an object person with prescribed service predicts a result of a prescribed service of the object person on the basis of actual data indicating situations of the prescribed service and non-linear prediction model, calculates a service providing method for providing the object person with services on the basis of a service providing capacity for providing the prescribed service and the predicted result, predicts virtual data indicating a change in situation of the object person when the prescribed service is provided to the object person according to the calculated service providing method, compares a first prediction result obtained by inputting the actual data to the non-linear prediction model with a second prediction result obtained by inputting the virtual data to the non-linear prediction model, and regulates the calculated service providing method.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a system and method for coordinating service delivery practices. [Background technology]

[0002] There are techniques that can effectively intervene in people and lead to behavioral changes in the target person. An example of human intervention is health guidance provided by insurers to insured persons. Health guidance involves encouraging insured persons to undergo specific health checkups and providing health guidance with the aim of promoting the health of insured persons or stabilizing operations by optimizing medical or nursing care costs.

[0003] Health guidance and other health projects provide health guidance to people who meet certain conditions, based on the resources or services that the business operator has available. For example, if a person's BMI (Body Mass Index), one of the test items in a specific health checkup, is below the normal value for their age, they will be advised that they are at risk of malnutrition and need to improve their lifestyle.

[0004] However, since there is a limit to the number of people who can actually receive intervention, it is not realistic to provide intervention to everyone who needs it. Therefore, in order to implement health care services by insurers in an effective and efficient manner using resources, a system that supports the selection of targets and intervention methods is desired.

[0005] Patent Document 1 discloses a method for assisting insurers in selecting recipients of intervention. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Japanese Patent Application Laid-Open No. 2009-151404 Summary of the Invention [Problem to be solved by the invention]

[0007] For example, when an insurer provides health guidance to insured persons, the improvement effect of the guidance on the target person is likely to vary depending on the health status of the person receiving the guidance (physical examination data, diagnosis of injury or illness, etc.). Therefore, it is important to analyze the health status of each individual target person before deciding whether to intervene. However, in the past, such individual analysis had not been established, and there was also a lack of evidence to use for judgment, so no consideration was given to methods for extracting targets who are more likely to benefit from intervention.

[0008] Therefore, the present invention provides a system and method for adjusting a service provision method that is easier to use. [Means for solving the problem]

[0009] In order to solve the above problem, in one aspect of the present invention, a service provision method adjustment system that adjusts the method of providing a specified service to a target person predicts the target person's results regarding the specified service based on actual data indicating the target person's status regarding the specified service and a non-linear prediction model, calculates a service provision method that is a method of providing the service to the target person based on the service provision capacity for providing the specified service and the predicted results, predicts virtual data that indicates changes in the target person's status when the specified service is provided to the target person according to the calculated service provision method, and compares a first prediction result obtained by inputting actual data into the non-linear prediction model with a second prediction result obtained by inputting virtual data into the non-linear prediction model, and adjusts the calculated service provision method. [Effects of the Invention]

[0010] According to the present invention, the service provision method can be adjusted in consideration of the predicted results when a predetermined service is provided to a target person. [Brief explanation of the drawings]

[0011] [Figure 1]1 is an overall configuration diagram of a service provision method adjustment system according to an embodiment; [Figure 2] FIG. 2 is a functional configuration diagram of an intervention adjustment flag estimation system. [Figure 3] FIG. 1 is a functional configuration diagram of an intervention history management system. [Figure 4] FIG. 2 is a functional configuration diagram of an intervention adjustment rule management system. [Figure 5] FIG. 1 is a functional configuration diagram of an intervention coordination system. [Figure 6] FIG. 10 is a diagram showing an actual data table of intervention subjects. [Figure 7] FIG. 10 is a diagram showing a prediction result table for intervention targets. [Figure 8] FIG. 10 is a diagram showing an intervention project information table. [Figure 9] FIG. 10 is a diagram showing an intervention method recommendation table. [Figure 10] FIG. 10 is a diagram showing a table of items improved by intervention. [Figure 11] FIG. 10 is a diagram showing an actual data table of intervention subjects. [Figure 12] FIG. 10 is a diagram showing a hypothetical data table for predicting post-improvement status of an intervention subject. [Figure 13] FIG. 10 is a diagram illustrating a prediction result comparison table. [Figure 14] FIG. 10 is a diagram showing an intervention adjustment criteria table. [Figure 15] FIG. 10 is a diagram illustrating a prediction result comparison table. [Figure 16] FIG. 10 is a diagram showing an intervention adjustment criteria table. [Figure 17] FIG. 10 is a diagram showing an intervention method recommendation table to which an intervention adjustment flag is assigned. [Figure 18] FIG. 10 is a diagram showing an intervention history data table. [Figure 19] FIG. 10 is a diagram showing a health information history data table. [Figure 20] FIG. 10 is a diagram showing an intervention evaluation table. [Figure 21] FIG. 10 is a diagram illustrating a behavioral change evaluation rule definition table. [Figure 22]FIG. 10 is a diagram showing a table that defines rules for evaluating the degree of improvement in health status. [Figure 23] FIG. 10 is a diagram showing an intervention adjustment rule table. [Figure 24] FIG. 10 is a diagram showing an adjusted intervention method recommendation data table. [Figure 25] 10 is a flowchart showing the process from receiving a user's input via a user interface to presenting an intervention method for each subject. [Figure 26] 10 is a flowchart showing a process related to intervention coordination. [Figure 27] FIG. 10 is a diagram showing a main screen of the UI of the service provision method adjustment system. [Figure 28] FIG. 10 is a diagram showing a screen for registering an intervention history. [Figure 29] FIG. 10 is a diagram showing a screen for registering information about an intervention project. [Figure 30] FIG. 10 is a diagram showing a screen for registering detailed information about an intervention project. [Figure 31] FIG. 10 is a diagram showing a screen for registering an evaluation rule. [Figure 32] FIG. 10 is a diagram showing a screen for registering information on intervention coordination. [Figure 33] FIG. 10 is a diagram showing a screen for presenting adjustment results for each intervention subject. DETAILED DESCRIPTION OF THE INVENTION

[0012] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. In the service provision method adjustment system of this embodiment, an intervention provider that provides intervention services (predetermined services) to people extracts a group of subjects at risk according to standard criteria, and then adjusts the subjects who are appropriate as recipients of a specific intervention and the intervention method.

[0013] In this embodiment, the predetermined service will be described by taking as an example an intervention service for people such as health guidance, etc. The predetermined service is not limited to a health care service such as health guidance, but may also be, for example, a coaching service related to exercise habits or study habits, an advice service related to well-being, or a life reconstruction support service for families receiving welfare assistance.

[0014] As will be described later, the service provision method adjustment system 1000 can utilize health information data (actual data) that indicates the health status and lifestyle habits of each individual, such as data from specific health checkups, and has a nonlinear prediction model that is trained using this health information data as data to be analyzed.

[0015] The service provision method coordination system 1000 generates a prediction result data table 110 (see FIG. 7) for intervention candidates from an actual data table 100 (see FIG. 6) for intervention candidates based on the medical examination history, etc., of the intervention target group and a non-linear prediction model (stored in the prediction model storage unit of FIG. 2). The service provision method coordination system 1000 generates an intervention method recommendation table 130 (see FIG. 9) for each intervention target person based on an intervention business information table 120 (see FIG. 8), and generates post-improvement virtual data 160 (see FIG. 12) for the intervention candidate from the actual data 150 (see FIG. 11) of the intervention candidate and a table 140 of improvement items due to intervention (see FIG. 10).

[0016] The service provision method adjustment system 1000 generates a prediction result comparison table 170 (see Figure 13) from the prediction results of the nonlinear prediction model, and assigns an intervention adjustment flag 228 (see Figure 17) to the intervention method recommendation table 220 (see Figure 17) for each intervention target based on the intervention adjustment standard rule table 180 (see Figure 14).

[0017] The service provision method coordination system 1000 generates an intervention evaluation data table 300 (see Figure 20) based on an intervention history data table 230 (see Figure 18), a health information history table 240 (see Figure 19), a behavioral change evaluation rule definition table 310 (see Figure 21), and a health status improvement degree evaluation rule definition table 320 (see Figure 22).

[0018] The service provision method adjustment system 1000 presents the results of adjusting the intervention method for each intervention target based on the intervention method recommendation table 220 for each intervention target to which an intervention adjustment flag has been assigned and the intervention adjustment rule table 330 (see Figure 23).

[0019] According to the service delivery method coordination system 1000, it is possible to estimate, from a group of subjects extracted as at risk, subjects who are expected to benefit from intervention and subjects who are not expected to benefit from intervention. An intervention provider can use the service delivery method coordination system 1000 to coordinate intervention subjects and intervention methods, thereby selecting an intervention method that can improve cost-effectiveness and improving user usability. The user here is an intervention provider who uses the service delivery method coordination system 1000.

[0020] The service provision method adjustment system 1 may have the following configuration, for example: The following configuration example does not limit the scope of the present invention.

[0021] (Configuration 1) A service provision method adjustment system that estimates an intervention method for an intervention target based on the estimated effect of the intervention, adjusts the intervention method, and presents the intervention target and the intervention method, and holds an actual data table regarding health information for each intervention candidate, a nonlinear prediction model that predicts the target's risk level, an intervention project information table that defines intervention projects that can be provided by the user, an intervention improvement item table that associates the intervention project with the intervention content, an intervention adjustment criteria table, and an intervention adjustment rule table, and holds records in the actual data table regarding health information for each intervention candidate, the prediction results by the nonlinear prediction model, and the intervention project information table. The intervention method for each subject is estimated from the data, and virtual data after improvement due to the intervention is generated from the records in the actual data table regarding the health information of each intervention candidate and the table of improvement items due to the intervention.A prediction result comparison table is generated based on the prediction results when the actual data regarding the health information of each intervention candidate and the virtual data after improvement due to the intervention are each input into a non-linear prediction model, and an intervention adjustment flag is assigned to each intervention subject from the prediction result comparison table and the intervention adjustment criteria table.The intervention method corresponding to the intervention subject is adjusted and presented based on the intervention project information table, the intervention adjustment flag assigned to each intervention subject, and the intervention adjustment rule.

[0022] (Configuration 2) In Configuration 1, by using XAI (Explainable artificial intelligence) in a nonlinear prediction model, it is possible to estimate the effect at the item level between actual data and virtual data after improvement through intervention.

[0023] (Configuration 3) In configuration 1, when adjusting an intervention method corresponding to an intervention target, there is a step of evaluating the degree of behavioral change due to the intervention based on the intervention history data and health information data.

[0024] (Configuration 4) In configuration 1, when adjusting an intervention method corresponding to an intervention target, there is a step of evaluating the degree of improvement in health condition due to the intervention based on the intervention history data and health information data.

[0025] (Configuration 5) In configuration 1, the intervention coordination rule table includes, as an item, whether or not there is cooperation with the intervention based on the intervention history data.

[0026] (Configuration 6) In configuration 1, the intervention adjustment rule table has, as items, an evaluation value of behavioral change in an intervention and an evaluation value of the degree of improvement in health status in an intervention, based on intervention history data and health information data.

[0027] (Configuration 7) In configuration 1, when adjusting the intervention method corresponding to the intervention target, there is a step of adjusting the intervention method corresponding to the intervention target based on the intervention adjustment rules and definition information regarding the intervention means, intervention frequency, and intervention time related to the intervention project.

[0028] (Configuration 8) In configuration 1, regarding the intervention adjustment rule, there is a step of estimating and proposing a new intervention adjustment rule by recording the processing performed by the user as a history as a result of adjusting the intervention method corresponding to the intervention target.

[0029] The service provision method coordination system 1000 of this embodiment can determine which people in a group that are determined to be at risk based on specific conditions should receive priority intervention, and present this to the user.

[0030] For example, if two subjects show similar trends in test items suggesting a risk of malnutrition, and health guidance on preventing malnutrition brings the test item closer to normal, it cannot be said that the risk was reduced equally according to the amount of change observed in the test item. This indicates that the two subjects' other test items differ, resulting in different health conditions and therefore different levels of risk for each item. Conventional methods have assumed that such perspectives are viewed only in terms of the risk item or as a general trend, and therefore do not take into account the individual's different risk levels. In contrast, the service delivery method adjustment system 1000 can estimate the intervention effect for each intervention subject based on the items expected to improve through the intervention and adjust the intervention method. [Example]

[0031] One embodiment will be described below with reference to Figures 1 to 33. This embodiment is an example in which, as part of an administrative initiative by a local government, when administrative public health nurses implement health guidance intervention projects for subjects at risk of transitioning to a state requiring nursing care in the future, it is possible to present to the administrative public health nurses an intervention method tailored for each intervention target, based on an estimated improvement effect of the intervention for each individual target, for a group of targets identified as at risk according to standard criteria.

[0032] First, definitions used in this specification will be explained.

[0033] "Intervention" refers to an approach taken by a subject to solve a problem, and does not include any coercive action taken against the subject. For example, a public health nurse provides health guidance to a subject on proper eating habits. This is an approach to correct the subject's irregular eating habits, and is included in the category of "intervention." However, it is not considered "intervention" until this approach forces the subject to change their behavior.

[0034] A "nonlinear predictive model" refers to a model that has learned about the relationship between cause and effect when the cause does not produce a proportional result. For example, the lifestyle habit of eating a late-night snack three or more times a week and the progression to a state requiring nursing care are thought to have a cause and effect relationship, but whether increasing the number of times from three to four or five times has a proportional impact on the outcome varies from person to person, and is not a proportional impact that is seen in all people. A "nonlinear predictive model" refers to a predictive model that has learned about the relationship between cause and effect in such cases, taking into account other causes that differ from person to person.

[0035] First, the configuration of a service provision method adjustment system 1000 according to an embodiment will be described with reference to FIGS.

[0036] 1, the service provision method adjustment system 1000 of this embodiment is configured such that an intervention adjustment flag estimation system 1, an intervention history management system 3, an intervention adjustment rule management system 4, and an intervention adjustment system 5 are connected via a communication network 2. The network 2 may be a LAN (Local Network) or a global network such as the Internet.

[0037] The functional configuration of each of the components 1, 3, 4, and 5 of the service providing method adjustment system 1000 will be described with reference to FIGS.

[0038] Each of the components 1, 3, 4, and 5 includes devices such as a processor 60, an input device 61, an output device 62, an auxiliary storage device 63, and a network interface 64, and these devices 60 to 64 are connected by a communication path 65.

[0039] The processor 60 is not limited to a CPU (Central Processing Unit) and may include a GPU (Graphics Processing Unit) or an ASIC (Application Specific Integrated Circuit). The processor 60 executes a computer program to realize each function processing unit. The input device 61 is a device through which a user inputs information to the service provision method adjustment system 1000, and includes, for example, a keyboard, a touch panel, a pointing device, and a voice instruction device. The output device 62 is a device that outputs information from the service provision method adjustment system 1000 to a user, and includes, for example, a monitor display, a printer, and a voice synthesizer. The auxiliary storage device 63 is a device with a relatively large storage capacity, such as a hard disk drive or an SSD (Solid State Drive). The network interface 64 is an interface for connecting to the communication network 2.

[0040] The functional configuration of the intervention adjustment flag estimation system 1 will be described using Figure 2. As shown in Figure 2, the intervention adjustment flag estimation system 1 can include, for example, the following functional units: a data acquisition processing unit 10, a prediction model input processing unit 11, a virtual data generation unit 12, an intervention method estimation unit 13, a user interface processing unit 14, an intervention adjustment flag estimation unit 15, an intervention improvement item storage unit 16, an intervention adjustment flag storage unit 17, an intervention project information storage unit 18, a health information storage unit 19, a prediction model storage unit 20, and an intervention adjustment standard storage unit 21. In the figure and in the following description, the user interface may be abbreviated as "UI."

[0041] The data acquisition processing unit 10 is a functional unit that performs processing to acquire necessary data. For example, if health information and a prediction model based on the health information are managed outside of the system 1000, the data acquisition processing unit 10 acquires the health information and the prediction model via the network 2 and stores them in the health information storage unit 19 and the prediction model storage unit 20.

[0042] The prediction model input processing unit 11 is a functional unit that generates data to be input to the prediction model from the health information data stored in the health information storage unit 19, and performs processing to input the data to the prediction model.

[0043] The virtual data generating unit 12 is a functional unit that refers to the information stored in the intervention improvement item storage unit 16 and generates post-improvement assumed data based on the actual data for each intervention target.

[0044] The intervention method estimation unit 13 is a functional unit that estimates an intervention method for each intervention target based on the prediction results obtained by inputting data into the prediction model and the information stored in the intervention project information storage unit 18.

[0045] The UI processing unit 14 is a functional unit that accepts and processes inputs of definition information stored in the intervention project information storage unit 18, definition information stored in the intervention improvement item storage unit 16, and definition information stored in the intervention adjustment standard storage unit 21 on the UI.

[0046] In the following explanation, the tables included in each storage unit (data store) will be described later. In this embodiment, it is assumed that a health information database and a prediction model that makes predictions based on health information exist outside the service provision method adjustment system 1000. However, this is not limited to this, and the health information database and the prediction model based on health information may be provided within the service provision method adjustment system 1000.

[0047] The functional configuration of the intervention history management system 3 will be described using Fig. 3. As shown in Fig. 3, the intervention history management system 3 can include, for example, a data recording unit 30, a data extraction unit 31, a UI processing unit 32, and an intervention history data storage unit 33.

[0048] The data recording unit 30 is a functional unit that generates and records intervention history data based on the content input on the UI. The data extraction unit 31 is a functional unit that searches for records under specific conditions in the intervention history data stored in the intervention history data storage unit 33 and extracts corresponding record groups. The UI processing unit 32 is a functional unit that accepts input of information related to the intervention history of each intervention target on the UI and passes it to the data recording unit 30.

[0049] The functional configuration of the intervention coordination rule management system 4 will be described using Fig. 4. As shown in Fig. 4, the intervention coordination rule management system 4 can include, for example, a data acquisition processing unit 40, an intervention coordination rule estimation processing unit 41, a UI processing unit 42, and an intervention coordination rule storage unit 43.

[0050] The data acquisition processing unit 40 is a functional unit that acquires data to be stored in the intervention adjustment rule storage unit 43. The intervention adjustment rule estimation processing unit 41 is a functional unit that estimates a new intervention adjustment rule based on the user's processing history regarding the estimation of the intervention adjustment rule and the recommendation of the intervention adjustment rule. The UI processing unit 42 is a functional unit that receives definition information of the intervention adjustment criteria and definition information of the intervention adjustment rule on the UI and stores them in the intervention adjustment criteria rule storage unit 43 and the processing history storage unit 44.

[0051] The functional configuration of the intervention coordination system 5 will be described using Fig. 5. As shown in Fig. 5, the intervention coordination system 5 is composed of, for example, an intervention history data acquisition processing unit 50, a health information data acquisition processing unit 51, an intervention evaluation data generation unit 52, an intervention coordination rule acquisition unit 53, an intervention method adjustment unit 54, a UI processing unit 55, an intervention evaluation data storage unit 56, a health information storage unit 57, a behavior change evaluation rule storage unit 58, and a health condition improvement degree evaluation rule storage unit 59.

[0052] The intervention history data acquisition processing unit 50 has a function to acquire data showing the history of interventions made by the intervention provider to the intervention target. The health information data acquisition processing unit 51 has a function to acquire data (such as health check data) showing the health status of the intervention target. The intervention evaluation data generation unit 52 has a function to generate data for evaluating intervention for the intervention target. The intervention adjustment rule acquisition unit 53 has a function to acquire rules for adjusting the intervention method, such as whether or not to intervene. The intervention method adjustment unit 54 has a function to adjust the intervention method. The UI processing unit 55 has a function to exchange information with the user via the input device 61 and the output device 62.

[0053] The intervention evaluation data storage unit 56 is a data store that stores the intervention evaluation data generated by the intervention evaluation data generation unit 52. The health information storage unit 57 is a data store that stores the health information acquired by the health information data acquisition and generation unit 51. The behavior change evaluation rule storage unit 58 is a data store that stores rules for evaluating the behavior change of the intervention recipient. The health condition improvement degree evaluation rule storage unit 59 is a data store that stores rules for evaluating the degree of improvement in the health condition of the intervention recipient.

[0054] The data structure handled by the service providing method adjustment system of this embodiment will be described with reference to FIGS.

[0055] The actual data table 100 of intervention subjects shown in Figure 6 is a data table based on health information data for each intervention subject ID, and when input into the nonlinear prediction model, the data has been converted into a data format that the nonlinear prediction model can accept input.

[0056] As shown in Figure 6, the actual data table 100 has items (which can also be called items) such as an intervention subject ID 101, an item 102 indicating that a late-night snack is eaten three or more times a week, an item 103 indicating that a late-night snack is not eaten three or more times a week, an item 104 indicating that breakfast is skipped three or more times a week, an item 105 indicating that breakfast is not skipped three or more times a week, an item 106 indicating that the questionnaire was not answered, an item 107 indicating that BMI is 25.0 or more, an item 108 indicating that BM1 is less than 25.0 and greater than or equal to 20.0 (25.0 > BMI 20.0), and an item 109 indicating that the person has not undergone a specific health checkup.

[0057] In this example, for simplicity, only questionnaire information regarding dietary habits, such as "I have a late-night snack at least three times a week" and "I skip breakfast at least three times a week," and indicator information based on biometric test information, such as "BMI," are represented as part of the input data to the nonlinear prediction model.

[0058] In this embodiment, only questionnaire information on dietary habits such as "I eat a midnight snack three or more times a week" and "I skip breakfast three or more times a week" and indicator information based on biopsy information such as "BMI" are taken as health information, but the present system 1000 is not limited to these. For example, the health information may also include medical receipt data containing the results of the subject's (intervention subject's) hospital diagnosis, drug receipt data containing prescription status, etc.

[0059] An identifier that uniquely identifies this record is stored in the intervention subject ID 101. Items 102, indicating that a late-night snack is eaten three or more times a week, to 109, indicating that a specific health checkup has not been taken, are configured with explanatory variables that the nonlinear prediction model can accept as input and their values. In this embodiment, some of the explanatory variables that the nonlinear prediction model accepts as input are item 102, indicating that a late-night snack is eaten three or more times a week, to item 109, indicating that a specific health checkup has not been taken, and it is assumed that the values are two values, "applicable" or "not applicable," but the present system 1000 is not limited to this. For example, the values may be categorical data defined by three or more values, such as "A," "B," and "C," or numerical data.

[0060] The prediction result table 110 for intervention subjects shown in Figure 7 is a table of prediction results for each intervention subject ID obtained when the actual data table 100 for intervention subjects is input into a non-linear prediction model, and as shown in Figure 7, it has the items of intervention subject ID 111 and probability of risk of needing care 112.

[0061] The intervention target ID 111 stores an identifier that uniquely identifies this record, and is synonymous with the intervention target ID 101 in the intervention target actual data table 100. The care-requiring risk probability 112 stores a predicted value estimated based on the input data for each intervention target ID 111.

[0062] In this embodiment, the nonlinear prediction model is an example of a model in which the outcome is a prediction of a future transition to needing nursing care based on information for each subject, but this does not apply to the present system 1000. For example, if the outcome of the nonlinear prediction model used is a model in which the outcome is a prediction of the future onset of diabetes, the item name of the probability of needing nursing care risk 112 may be treated as the probability of developing diabetes.

[0063] 8 is a table for storing information on intervention projects that can be provided by a user, and as shown in Fig. 8, it is composed of, for example, an intervention method ID 121, an intervention project 122, an intervention content 123, an intervention means 124, an intervention frequency 125, an intervention time 126, an intervention resource 127, and a response rule 128. The intervention project information table 120 is stored in the intervention project information storage unit 18 of the intervention adjustment flag estimation system 1.

[0064] The intervention method ID 121 stores an identifier that uniquely identifies this record. The intervention project 122 stores the name of an intervention project that the user of this system 1000 can provide to each intervention target. The intervention content 123 stores an identifier related to the intervention effort. In this example, it is assumed that "health guidance" or "recommendation to receive medical examination" or the like is stored with reference to government efforts.

[0065] The intervention means 124 stores the means of intervention. In this embodiment, it is assumed that "(individual) visit," "attendance," "telephone call," "letter," etc. are stored, referring to the efforts of health guidance in government health programs, but this is not restrictive. The intervention frequency 125 stores the frequency of intervention. The intervention time 126 stores the time required for one intervention. The intervention resource 127 stores the number of people or number of cases to which the user can provide the intervention method uniquely identified by the intervention method ID 121. The correspondence rule 128 stores standard target selection rules for each intervention method.

[0066] 9 is a table that links an intervention method for each subject based on the nursing care need risk probability 112 in the intervention subject prediction result table 110 and the correspondence rule 127 in the intervention project information table 120. As shown in FIG. 9, the intervention method recommendation table 130 may include, for example, an intervention subject ID 131, a risk probability 132, an intervention project 133, an intervention content 134, an intervention means 135, an intervention frequency 136, and an intervention time 137.

[0067] The intervention target ID 131 stores an identifier that uniquely identifies this record, and is synonymous with the intervention target ID 101 in the intervention target actual data table 100. The risk probability 132 is a predicted probability based on a non-linear prediction model for the target, and is synonymous with the nursing care need risk probability 112 in the intervention target prediction result table 110. The intervention project 133 to intervention time 137 each store elements related to the intervention method, and are based on the information stored in the items of the same name in the intervention project information table 120.

[0068] 10 is a table that stores definition information of which item of the intervention target the intervention is aimed at solving or improving, corresponding to the intervention project and the intervention content, and as shown in Fig. 10, the improvement item table 140 can include, for example, an improvement item ID 141, an intervention project 142, an intervention content 143, and an improvement item 144. The intervention project information table 140 is stored in the intervention improvement item storage unit 16 of the intervention adjustment flag estimation system 1.

[0069] An identifier that uniquely identifies this record is stored in the improvement item ID 141. Elements related to the intervention method are stored in the intervention project 142 and intervention content 143, and are based on information stored in the items of the same name in the intervention project information table 120. The improvement item 144 corresponds to the intervention method and is the target of problem solving or improvement, and receives input information from the system user via the UI regarding item information corresponding to the explanatory variables defined in the actual data table 100 of the intervention target, and stores the information.

[0070] The actual data table 150 of intervention subjects shown in FIG. 11 corresponds to a record with subject ID "0001" extracted from the actual data table 100 of intervention subjects, in order to explain an example.

[0071] The post-improvement virtual data table 160 of the intervention recipient shown in Fig. 12 is a table generated based on the actual data table 150 of the intervention recipient, and contains data on the case where an intervention provided by a user to the intervention recipient improves an item corresponding to an improvement item defined in the table of improvement items by intervention 140. As shown in Figs. 11 and 12, the post-improvement virtual data table 160 has the same number of items and item names as the actual data table 150 of the intervention recipient.

[0072] In this example, as shown in Figures 11 and 12, only the record with the intervention subject ID 161 of "0001" will be described as an example. The item "having a late-night snack three or more times a week" is an item with an increased risk compared to the item "having a late-night snack not three or more times a week," so it is estimated that after improvement, the item "having a late-night snack not three or more times a week" will be applicable. In the record with the intervention subject ID 161 of "0001," "having a late-night snack not three or more times a week" was applicable in the actual data 150 of the intervention subject, so it is assumed that there will be no change in this item in the post-improvement virtual data 160 of the intervention subject. With regard to the item "skipping breakfast three or more times a week," this is an item that increases risk compared to the item "not skipping breakfast three or more times a week," and "skipping breakfast three or more times a week" is applicable in the actual data 150 of the intervention subject, so virtual data is generated with this item not applicable in the post-improvement virtual data 160 of the intervention subject. With regard to the item "BMI," the item "BMI≧25.0" is an item that increases risk compared to the item "25.0>BMI≧20.0," and "BMI≧25.0" is applicable in the actual data 150 of the intervention subject, so virtual data is generated with this item not applicable in the post-improvement virtual data 160 of the intervention subject. By repeatedly performing the same process for each intervention subject ID 161, values for each item in the post-improvement virtual data table 160 of the intervention subject are estimated, generated, and stored.

[0073] In this embodiment, determining whether a change in an improvement item results in an improvement for the intervention subject is realized by generating post-improvement virtual data and repeatedly comparing the results input to a non-linear prediction model, but the present system 1000 does not rely on this. For example, for BMI, a predefined condition that 25.0 > BMI ≥ 20.0 is considered a normal value may be accepted, and post-improvement virtual data may be generated based on the condition.

[0074] 13 is a table that combines the prediction results when the improved virtual data 160 of the intervention recipient is input into a non-linear prediction model with the prediction results in the intervention recipient prediction result table 110. As shown in Fig. 13, the prediction result comparison table 170 includes the item names of the intervention recipient ID 171, the probability of risk of needing care (actual data) 172, and the probability of risk of needing care (virtual data) 173.

[0075] The intervention subject ID 171 stores an identifier that uniquely identifies this record, and is synonymous with the intervention subject ID 101 in the intervention subject actual data table 100. The care-requiring risk probability (actual data) 172 corresponds to the predicted probability of a non-linear prediction model using the actual data of the subject, and is synonymous with the care-requiring risk probability 112 in the intervention subject prediction result table 110. The care-requiring risk probability (virtual data) 173 stores the predicted probability of a non-linear prediction model using the improved virtual data of the subject.

[0076] The intervention adjustment criteria table 180 shown in Fig. 14 is a table that contains criteria information for determining the range of intervention targets for which an intervention adjustment flag indicating a target for adjusting the intervention method should be assigned, based on the estimated effect of the intervention. As shown in Fig. 14, the intervention adjustment criteria table 180 includes, for example, the item names of an evaluation target 181, an adjustment range 182, and an adjustment target 183.

[0077] The evaluation target 181 is an item that stores an index that is focused on when assigning an intervention adjustment flag. In the present embodiment, in the case of the difference in risk probability, it refers to the magnitude of the absolute value of the difference between the nursing care need risk probability (virtual data) 173 and the nursing care need risk probability (actual data) in the prediction result comparison table 170. The adjustment range 182 is an item that stores the numerical range of the index defined in the evaluation target 181. The adjustment target 183 is an item that stores additional conditions for the intervention target ID that belongs to the numerical range defined in the adjustment range 182.

[0078] This embodiment is an example of a case where the effect is estimated from the difference in predicted probability values in a nonlinear prediction model, but the present system 1000 does not rely on this. For example, XAI may be used as a nonlinear prediction model, and the improvement effect may be estimated for each improvement item rather than the overall risk probability, and an intervention adjustment flag may be assigned.

[0079] This modified example will be explained using Figures 15 and 16. The prediction result comparison table 190 in this modified example is a table in which the actual data 150 of the intervention target and the post-improvement virtual data 160 of the intervention target are input into a non-linear prediction model, and the weight of each item (which can be regarded as the degree of influence on the risk probability) is extracted by XAI.

[0080] The prediction result comparison table 190 includes, for example, an intervention subject ID 191, a probability of needing nursing care risk 192, an item 193 indicating that a late-night snack is eaten three or more times a week, an item 194 indicating that a late-night snack is not eaten three or more times a week, an item 195 indicating that breakfast is skipped three or more times a week, an item 196 indicating that breakfast is not skipped three or more times a week, an item 197 indicating that no response was made to the questionnaire, an item 198 indicating that BMI is 25.0 or more, an item 199 indicating that 25.0 > BMI 20.0, and an item 200 indicating that the person has not undergone a specific health checkup.

[0081] The intervention recipient ID 191 stores an identifier that uniquely identifies this record and is synonymous with the intervention recipient ID 101 in the intervention recipient actual data table 100. The nursing care need risk probability (actual data) 192 corresponds to the predicted probability of a nonlinear prediction model using the actual data of the subject, and is synonymous with the nursing care need risk probability 112 in the intervention recipient prediction result table 110. Each item, from item 193 indicating having a midnight snack three or more times a week to item 200 indicating not having undergone a specific health checkup, stores a weight output by XAI for each item, and these weights indicate the contribution of the value of the corresponding item to the overall predicted value, the nursing care need risk probability 192. Here, the sign of the value indicates the contribution in the direction of an increase in the predicted value if positive, or the contribution in the direction of a decrease if negative, and the value indicates the magnitude of the contribution. In this example, a positive sign and a larger value indicate an increase in the nursing care need risk, and a negative sign and a larger value indicate a decrease in the nursing care need risk.

[0082] The modified intervention adjustment criteria table 210 shown in Figure 16 is a table containing criteria information for determining, based on the estimated effectiveness of the intervention, the range of intervention targets for which an intervention adjustment flag indicating the target for adjusting the intervention method should be assigned, and includes the item names of evaluation target 211, adjustment range 212, and adjustment target 213.

[0083] The evaluation target 211 is an item that stores an index to be focused on when assigning an intervention adjustment flag. In the case of the prediction result comparison table 190 using XAI, it refers to the magnitude of the difference in weight for each item. The adjustment range 212 is an item that stores the numerical range of the index defined in the evaluation target 211. The adjustment target 213 is an item that stores additional conditions for the intervention target ID that belongs to the numerical range defined in the adjustment range 212.

[0084] The intervention method recommendation table 220 shown in Figure 17 is a table in which an intervention adjustment flag item has been added to the intervention method recommendation table 130, which corresponds intervention targets and intervention methods based on the intervention target prediction result table 110 and the intervention project information table 120.

[0085] As shown in FIG. 17, the intervention method recommendation table 220 includes an intervention target ID 221 , a risk probability 222 , an intervention project 223 , an intervention content 224 , an intervention method 225 , an intervention frequency 226 , an intervention time 227 , and an intervention adjustment flag 228 .

[0086] In the intervention adjustment flag 228, a flag is stored for each record of the intervention subject ID corresponding to a high intervention effect and a flag for each record of the intervention subject ID corresponding to a low intervention effect, based on the adjustment target 183 in the intervention adjustment criteria table 180.

[0087] The intervention history table 230 shown in Figure 18 is a table that, for each intervention subject ID, shows the intervention methods implemented in the past, their date information, and whether or not the subject cooperated with the intervention, and includes the following items: intervention history ID 231, intervention subject ID 232, intervention project 233, intervention content 234, intervention method 235, intervention frequency 236, intervention time 237, intervention start date 238, intervention end date 239, and whether or not the subject cooperated with the intervention 240.

[0088] The intervention history ID 231 stores an identifier that uniquely identifies this record. The intervention target ID 232 is synonymous with the intervention target ID 101 in the intervention target actual data table 100. The items from intervention project 233 to intervention time 237 are the same as the items with the same names in the intervention project information table, and information on the intervention method actually implemented for the intervention target ID is received and stored via a UI from the user. The intervention start date 238 stores the date and time when the intervention method started. The intervention end date 239 stores the date and time when the intervention method ended. The intervention cooperation presence / absence 240 stores whether or not the intervention target cooperated when the intervention was implemented. For example, if the intervention was conducted by phone and the person answered, the intervention cooperation is recognized as having been implemented, whereas if the call was not connected or was immediately hung up, the intervention cooperation is recognized as not having been implemented.

[0089] The health information history data table 240 shown in Figure 19 is a table of health information history data for each intervention subject ID for a specific year and month in the past, and includes the following items: resident ID 241, item 242 indicating that the person eats a late-night snack or snack three or more times a week, item 243 indicating that the person skips breakfast three or more times a week, item 244 indicating that the person has not responded to the questionnaire, BMI 245, item 246 indicating that the person has not undergone a specific health checkup, and recording date and time 247.

[0090] An identifier that uniquely identifies this record is stored in the resident ID 241. Items from item 242 indicating that the person has a midnight snack three or more times a week to item 246 indicating that the person has never undergone a specific health checkup are data related to the resident's health information, and are examples in this embodiment. The recording date and time 247 stores the date and time when the record was acquired or recorded.

[0091] 20 is a table that stores, for each intervention history ID and intervention recipient ID, an evaluation value of the effect of past interventions on the intervention recipient, and includes the following items: intervention history ID 301, intervention recipient ID 302, whether or not there was cooperation with the intervention 303, behavioral change 304, and degree of improvement in health condition 305. The item names and evaluation axes in this embodiment are merely examples, and the present system 1000 does not rely on them.

[0092] The intervention history ID 301 is a unique identifier, synonymous with the intervention history ID 231 in the intervention history table 230. The intervention recipient ID 302 is a unique identifier, synonymous with the intervention recipient ID 101 in the actual data table 100 for intervention recipients. The item 303 indicating whether or not there was cooperation with the intervention is synonymous with the item 240 indicating the presence or absence of an intervention history in the intervention history table 230. The behavioral change 304 stores an estimated evaluation value based on the data of the improvement items in the health information history data table 240 and the behavioral change evaluation rules defined in the behavioral change evaluation rule definition table 310. The health condition improvement level 305 stores an estimated evaluation value based on the data of the improvement items in the health information history data table 240 and the health condition improvement level evaluation rules defined in the health condition improvement level evaluation rule definition table 320.

[0093] 21 shows a behavior change evaluation rule definition table 310 that stores rules and evaluation values for evaluating the effectiveness of an intervention based on past health information history data for each intervention recipient. This table 310 includes, for example, items such as a behavior change evaluation rule ID 311, an evaluation criterion 312, and an evaluation value 313.

[0094] The behavior change evaluation rule ID 311 stores an identifier that uniquely identifies this record. The evaluation criteria 312 stores the condition definitions used for evaluation. The evaluation value 313 stores the evaluation value. The evaluation criteria 311 and evaluation value 312 are examples, and are not limited to these and can be adjusted to suit the user's purpose and evaluation criteria. The same applies to the tables described below.

[0095] 22 shows a health condition improvement degree evaluation rule definition table 320, which stores rules and evaluation values for evaluating the effectiveness of interventions based on past health information history data for each intervention recipient. This table 320 includes, for example, items such as a health condition improvement degree evaluation rule ID 321, an evaluation criterion 322, and an evaluation value 323.

[0096] An identifier that uniquely identifies this record is stored in the health condition improvement evaluation rule ID 321. A condition definition for evaluation is stored in the evaluation criteria 322. An evaluation value is stored in the evaluation value 323.

[0097] 23 is a table that stores definition information of the rule of intervention coordination for a record that links an intervention target to which an intervention coordination flag is assigned and an intervention method. This table 330 includes, for example, items such as an intervention coordination rule ID 331, a condition 332, an intervention coordination item 333, and an adjustment method 334.

[0098] An identifier that uniquely identifies this record is stored in the intervention adjustment rule ID 331. The condition 332 stores judgment logic using the evaluation values of intervention cooperation 303, behavior change 304, and health status improvement level 305 in the intervention evaluation table 300, and the intervention adjustment flag as elements. The intervention adjustment item 333 stores a definition of which element in the intervention to adjust when the condition 332 is met. The adjustment method 334 stores information on the adjustment method for the intervention adjustment item 333 when the condition 332 is met.

[0099] The adjusted intervention method rule table 340 shown in Figure 24 is a table after adjusting the records of recommended intervention methods for records in the intervention method recommendation table 220 shown in Figure 17 that link intervention targets and intervention methods to which an intervention adjustment flag has been assigned, based on the intervention adjustment rule table 330 shown in Figure 23, the input items to the user interface described in Figure 30 (intervention means, intervention frequency, intervention time), and definition information of the costs required for implementation.

[0100] The basic structure of this table 340 is the same as the intervention method recommendation table 220 described in Fig. 17. This table 340 includes, for example, the following items: intervention target ID 341, risk probability 342, intervention project 343, intervention content 344, intervention method 345, intervention frequency 346, intervention time 347, and intervention adjustment flag 348.

[0101] FIG. 24 shows an example of intervention method recommendation data after adjustment. For example, the record with the intervention target ID "0001" has the intervention adjustment flag 348 set, and is considered to have a high intervention effect. To determine the rule to be adopted as the adjustment method for this record, a search is made for matching conditions in the conditions 332 of the intervention adjustment rule table 330. In this example, the record with the intervention adjustment rule ID "1" is found to match, and the intervention adjustment item 333 and adjustment method 334 of that record are referenced. As a result, the adjustment method of continuing the previous intervention method matches the intervention method that has already been recommended, so the adjustment is made to not change the existing recommendation.

[0102] As a different example, the record with the intervention target ID "0003" has the intervention adjustment flag 348 set, and the intervention effect is considered to be low. To determine the rule to be used as the adjustment method for this record, a matching condition is searched for among the conditions 332 in the intervention adjustment rule table 330. In this example, the record with the intervention adjustment rule ID "3" is matched, and the intervention adjustment item 333 and adjustment method 334 of that record are referenced. The adjustment method corresponds to changing the values of the intervention frequency and intervention time items from the standard values to values that reduce costs, while using past intervention methods as a base. In this system 1000, the intervention frequency definition table 740 and intervention time definition table 750 shown in FIG. 30 are referenced, and after adjusting each, the corresponding intervention frequency and intervention time are determined. In this example, the standard value for intervention frequency is "once every three months," but in order to reduce costs, the adjusted value is "once every six months." The standard value for intervention time is "more than one hour but less than two hours," but in order to reduce costs, the adjusted value is "more than 30 but less than one hour."

[0103] As another example, the record with the intervention target ID "0004" has the intervention adjustment flag 348 set, indicating a high intervention effect. To determine the rule to be used as the adjustment method for this record, a matching condition is searched for in the conditions 332 of the intervention adjustment rule table 330. In this example, the record with the intervention adjustment rule ID "2" is matched, and the intervention means 332, intervention adjustment item 333, and adjustment method 334 of the record are referenced. The adjustment method corresponds to increasing the value of the intervention means item by a cost increase and setting the values of the intervention frequency and intervention time items to their standard values. The intervention means definition table 730 in FIG. 30 is referenced to determine the intervention means corresponding to each adjustment. In this example, the standard value for the intervention means is "daytime care," but since this adjustment increases costs, the adjustment is changed to "visit" after adjustment. At this time, it is also determined whether the intervention resource fits within the intervention resources 127 defined in the intervention project information table 120 shown in FIG. 8, and adjustment is made only if adjustment is possible.

[0104] This example is for use in a project related to nursing care prevention, but is not to be construed as limiting. Although "intervention method," "intervention frequency," and "intervention time" are listed as items to be adjusted, other adjustments such as "intervention time" (spring or summer) or "way of speaking or interacting" (speaking slowly, listening attentively) may also be incorporated.

[0105] The processing performed by the service providing method adjustment system 1000 will be described with reference to FIGS.

[0106] The main flow executed by the service providing method adjustment system 1000 will be described with reference to FIG.

[0107] When information on a subject group as a candidate for intervention, information on a nonlinear prediction model, an estimate of an individual intervention method for that subject group, and a request for adjustment are received from user operations on the UI, these are notified to the data acquisition processing unit 10 of the intervention adjustment flag estimation system 1.

[0108] The data acquisition processing unit 10 of the intervention adjustment flag estimation system 1 stores the information received from the user through the UI as new health information and a nonlinear prediction model in the health information storage unit 19 and prediction model storage unit 20 of the intervention adjustment flag estimation system 1 (step 600), and then proceeds to step 601.

[0109] The prediction model input processing unit 11 of the intervention adjustment flag estimation system 1 acquires the nonlinear prediction model stored in the prediction model storage unit 20 and the health data of the subject group stored in the health information storage unit 19, inputs them into the nonlinear prediction model, acquires the prediction result table 110 of intervention subjects which is the prediction result (step 601), and proceeds to step 602.

[0110] The intervention method estimation unit 13 of the intervention adjustment flag estimation system 1 receives the prediction result table 110 of the intervention target person obtained in the previous section, obtains the intervention project information table 120 from the intervention project information storage unit 18 of the intervention adjustment flag estimation system 1, and generates the intervention method recommendation table 130 by matching the appropriate intervention method for each intervention target ID based on the intervention resource 127 and the corresponding rule 128 (step 602), and then proceeds to step 603.

[0111] The virtual data generation unit 12 of the intervention adjustment flag estimation system 1 acquires the table of improvement items due to intervention 140 stored in the storage unit for improvement items due to intervention, and refers to the intervention project 133 and intervention content 134 in the intervention method recommendation table 130 generated in the previous section, and searches for records in which synonymous data is stored in the intervention project 142 and intervention content 143 in the table of improvement items due to intervention 140.

[0112] The intervention adjustment flag estimation system 1 refers to the data of the improvement item 144 in the record searched in the previous section, and searches for the corresponding item in the actual data 150 of the intervention target.

[0113] The intervention adjustment flag estimation system 1 searches for a combination of data for the relevant item that will work to reduce the prediction probability of the nonlinear prediction model, generates that data combination as post-improvement virtual data 160 for the intervention target (step 603), and proceeds to step 604.

[0114] The prediction model input processing unit 12 of the intervention adjustment flag estimation system 1 inputs the improved virtual data 160 of the intervention target person generated in the previous section into a non-linear prediction model, and generates a prediction result comparison table 170 by combining the prediction result with the prediction result table 110 of the intervention target person (step 604), and then proceeds to step 605. The intervention adjustment flag estimation unit 15 of the intervention adjustment flag estimation system 1 refers to the intervention adjustment criteria table 180 in the intervention adjustment criteria storage unit 21, generates an intervention method recommendation table 220 in which an intervention adjustment flag is added as a new item to the intervention method recommendation table 130, stores it in the intervention adjustment flag storage unit 17 (step 605), and proceeds to step 606.

[0115] The data extraction unit 31 of the intervention history management system 3 receives the intervention method recommendation table 220 described above, searches for the data stored in the intervention history data storage unit 22 based on the intervention target ID contained therein, extracts the intervention history table 230 (step 606), and proceeds to step 607.

[0116] The intervention history data acquisition processor 50 of the intervention coordination system 5 acquires the intervention history table 230 generated in the previous section. The health information data acquisition processor 51 of the intervention coordination system 5 queries the data server storing health information based on the intervention target ID 231, membership start date 238, intervention end date 239, and intervention cooperation status 240 in the intervention history table 230, extracts the data, and then generates the health information history data table 240 (step 607), stores it in the health information storage unit 57, and proceeds to step 608. The intervention adjustment flag estimation system 1 adjusts the intervention method for each target (step 608), and presents a proposal for adjusting the intervention method to the user via the UI (step S609).

[0117] A flow for adjusting an intervention method for each subject will be described with reference to Figure 26. This intervention method adjustment flow corresponds to step 608 in Figure 25.

[0118] The data acquisition processing unit 50 of the intervention coordination system 5 acquires the intervention method recommendation table 220 described in the previous section, refers to the intervention coordination flag 228, searches for records of the intervention target ID to which the intervention coordination flag is assigned, repeats the intervention coordination flow for the number of corresponding records (step 610), and proceeds to step 611.

[0119] The intervention coordination system 5 searches for the intervention target ID based on the intervention target ID 231 in the intervention history data table 230 (step 611). If the search result shows that there is a record with the corresponding intervention target ID (step 612: Y), the intervention coordination system 5 proceeds to step 613, and if there is no record with the corresponding intervention target ID (step 612: N), the intervention coordination system 5 proceeds to step 619.

[0120] The intervention evaluation data generation unit 52 of the intervention coordination system 5 references the intervention cooperation presence / absence 240 item in the intervention history data table 230, acquires a record that indicates "Yes" (step 613), and proceeds to step 614.

[0121] The health information data acquisition processing unit 51 of the intervention coordination system 5 references the data for intervention start date 238 and intervention end date 239 of the record obtained in the previous section in intervention history data table 230, and acquires the corresponding data in the health information storage unit 57 (step 614). Here, data recorded immediately before the date of the data stored in intervention start date 238 and data recorded after the date of the data stored in intervention end date 239 are searched for, and health information history data table 240 is acquired. The intervention coordination system 5 then proceeds to step 615.

[0122] The intervention evaluation data generation unit 52 of the intervention coordination system 5 acquires the behavior change evaluation rule table 310 stored in the behavior change evaluation rule storage unit 58, searches for data items that satisfy the evaluation criteria 312 defined in the behavior change evaluation rule table 310 in the health information history data table 240 obtained by the process in the previous section, and estimates the behavior change evaluation value 313 for each intervention target ID (step 615).The intervention coordination system 5 then proceeds to step 616.

[0123] The intervention evaluation data generation unit 52 of the intervention coordination system 5 acquires the health condition improvement degree evaluation rule table 320 stored in the health condition improvement degree evaluation rule storage unit 59, searches the health checkup history data table 240 obtained by the processing in the previous section for data items that satisfy the evaluation criteria 322 defined in the health condition improvement degree evaluation rule definition table 320, and estimates the evaluation value 323 regarding the health condition improvement degree for each intervention target ID (step 616). Then, the intervention coordination system 5 proceeds to step 617. The intervention evaluation data generation unit 52 of the intervention coordination system 5 compares the evaluation value regarding behavioral change and the evaluation value regarding the degree of health status improvement for each intervention target ID to generate an intervention evaluation data table 300 as a single record, stores it in the intervention evaluation data storage unit 55 (step 617), and proceeds to step 618.

[0124] The data acquisition processing unit 40 of the intervention coordination rule management system 4 acquires the intervention coordination rule definition table 330 stored in the intervention coordination rule storage unit 43. Furthermore, the data acquisition processing unit 40 of the intervention coordination rule management system 4 searches for and acquires records having a value in the intervention coordination flag item 228 in the intervention method recommendation table 220. The data acquisition processing unit 40 of the intervention coordination rule management system 4 acquires the intervention evaluation table 300 from the intervention evaluation data storage unit 55 of the intervention coordination system 5.

[0125] The data acquisition processing unit 41 of the intervention adjustment rule management system 4 refers to the conditions 332 in the intervention adjustment rule definition table 330 and the intervention evaluation table 300 for records that have a value in the intervention adjustment flag item 228 in the intervention method recommendation table 220 described in the previous section, searches for and estimates the corresponding intervention adjustment rule ID (step 618), and proceeds to step 619.

[0126] The intervention method adjustment unit 54 of the intervention adjustment system 5 refers to the intervention adjustment item 333 and the adjustment method 334 in the intervention adjustment rule definition table 330, and adjusts the item stored in the intervention adjustment item 333 in the record in the intervention method recommendation table 220 using the method stored in the adjustment method 334 (step 619). This ends the intervention method adjustment flow, and the process proceeds to step 609 in FIG. 26.

[0127] The UI processing unit 55 of the intervention coordination system 5 presents the adjusted intervention method recommendation data table 340 on the UI (step 609).

[0128] The UI provided to the user by the service providing method adjustment system 1000 will be described with reference to FIGS.

[0129] 27, a main screen 800 in the UI of the service provision method adjustment system 1000 is a screen that accepts movement to each UI for registering intervention project information, intervention history, intervention evaluation rules, and intervention adjustment information that are required for estimating and adjusting an intervention method for an individual. The main screen 800 transitions to an input / registration screen for each piece of information in response to a user operation for registering intervention history 801, registering intervention project information 802, registering evaluation rules 803, or registering intervention adjustment information 804.

[0130] The main screen 800 accepts input 805 of a dataset for the target group for intervention method estimation and adjustment, and input 806 of a nonlinear prediction model to be used. By operating 807, the main screen 800 presents the estimation of the intervention method and adjustment results for each intervention target based on the various registered information and the target group dataset that have been accepted.

[0131] As shown in Figure 28, the intervention history input screen 700 is a screen that accepts input of the intervention project, intervention method, intervention start date, intervention end date, and whether or not intervention cooperation was provided, implemented for each user's intervention target ID. The screen 700 accepts input operations related to adding an intervention history record 701, editing an existing intervention history record 702, and deleting an existing intervention history record 703. By confirming input 704, the screen 700 stores the record accepted on the UI as the intervention project information table 120 in the intervention history data storage unit 33 of the intervention history management system 3.

[0132] 29, the input screen 710 for intervention project information is a screen that accepts input regarding intervention projects, intervention methods, intervention resources, and response rules that can be provided by the user. When the screen 710 accepts input operations regarding adding an intervention project information record 713, editing an existing intervention project information record 714, or deleting an existing intervention project information record 715, the record accepted on the UI is stored as the intervention history table 230 in the intervention project information storage unit 18 of the intervention adjustment flag estimation system 1 by input confirmation 716.

[0133] The input operation for Edit Improvement Item 711 enables input to a screen 720 for defining improvement items due to intervention, as shown in the lower part of Fig. 29, and can accept input of improvement items for the corresponding intervention method. The screen 720 accepts input operations for adding 721 records of intervention method IDs and improvement items, editing 722 records of improvement items corresponding to existing intervention method IDs, and deleting 723 records, and an input confirmation 724 causes the records accepted on the UI to be stored as the intervention improvement item table 140 in the intervention improvement item storage unit 16 of the intervention adjustment flag estimation system 1.

[0134] As shown in the upper part of Fig. 30 , inputting the intervention method details definition 712 enables input to an intervention method definition input screen 730, and can accept input of the element definition of the intervention method that can be provided by the user. The screen 730 accepts input operations for adding 731 an intervention method ID and an intervention method record, editing 732 an existing intervention method ID record, and deleting 733, and by confirming input 734, the record accepted on the UI is stored as intervention project information in the intervention project information storage unit 18 of the intervention adjustment flag estimation system 1.

[0135] Then, as shown in the center of Fig. 30 , an input screen 740 for intervention frequency definition becomes available for input, and input regarding element definitions of intervention methods that can be provided by the user can be accepted. The screen 740 accepts input operations regarding addition 741 of an intervention frequency ID and an intervention frequency record, editing 742 of an existing intervention method ID record, and deletion 743, and by confirming input 744, the record accepted on the UI is stored as intervention project information in the intervention project information storage unit 18 of the intervention adjustment flag estimation system 1.

[0136] 30, an input screen 750 for inputting intervention time definitions becomes available, and input regarding element definitions of intervention methods that can be provided by the user can be accepted. The screen 750 accepts input operations regarding addition 751 of an intervention time ID and an intervention time record, editing 752 of an existing intervention method ID record, and deletion 753, and by confirming input 754, the record accepted on the UI is stored as intervention project information in the intervention project information storage unit 18 of the intervention adjustment flag estimation system 1.

[0137] 31, when an input operation for registering an evaluation rule 803 is performed on the main screen 800, the screen transitions to screens 760 and 770. These screens accept input about the evaluation definition of an intervention according to the user's purpose and evaluation axis.

[0138] An input screen 760 for behavioral change evaluation rules, such as the one shown in the upper part of Figure 31, accepts input for the definition of behavioral change evaluation rules. The screen 760 accepts input operations for adding 761 a record of a behavioral change evaluation rule ID, evaluation criteria, and evaluation value, and editing 762 and deleting 763 an existing behavioral change evaluation rule ID record. The screen 760 stores the record accepted on the UI by confirming input 764 as a behavioral change evaluation rule definition table 310 in the behavioral change evaluation rule storage unit 58 of the intervention coordination system 5.

[0139] An input screen 770 for the assessment rule of the health condition improvement degree, as shown in the lower part of Fig. 31, accepts input for the assessment rule definition of the health condition improvement degree. The screen 770 accepts input operations for adding 771 a record of the health condition improvement degree assessment rule ID, and the assessment criteria and assessment value, and editing 772 and deleting 773 a record of an existing health condition improvement degree assessment rule ID. The screen 770 stores the record accepted on the UI by input confirmation 774 as the health condition improvement degree assessment rule definition table 320 in the health condition improvement degree assessment rule storage unit 59 of the intervention coordination system 5.

[0140] 32, when an input operation for registering intervention adjustment information 804 is performed on the main screen 800, the screen transitions to screens 780 and 790. These screens accept input of definitions related to intervention method adjustment according to the user's purpose.

[0141] An input screen 780 for intervention adjustment criteria, such as that shown in the upper part of Fig. 32, accepts input regarding the definition of the intervention adjustment criteria. The screen 780 accepts input operations regarding whether or not the criteria are applied, the evaluation target, the adjustment range, adding a record to be adjusted (781), editing an existing intervention adjustment criteria record (782), and deleting (783). The screen 780 stores the record accepted on the UI by confirming input (784) as the intervention adjustment criteria definition table 180 in the intervention adjustment criteria storage unit 21 of the intervention adjustment flag estimation system 1.

[0142] An intervention coordination rule input screen 790 such as that shown in the lower part of Fig. 32 accepts input for the intervention coordination rule definition. The screen 790 accepts input operations for adding 791 records of the intervention coordination rule ID, conditions, intervention coordination items, and adjustment methods, and for editing 792 and deleting 793 records of existing intervention coordination rule IDs. The screen 790 stores the records accepted on the UI by confirming input 794 as the intervention coordination rule definition table 330 in the intervention coordination rule storage unit 43 of the intervention coordination management system 4.

[0143] FIG. 33 shows a table 810 that defines the standard intervention method before adjustment, and a table 811 that defines the adjusted intervention method.

[0144] The recommendation screen for the intervention method after adjustment shown in Fig. 33 presents a recommendation table 810 for intervention recipients and intervention methods before adjustment, as well as a recommendation table 811 for intervention recipients and intervention methods after adjustment. By inputting the confirmation of adjustment rules 812, the intervention adjustment rule definition table 790 in Fig. 32 is also presented.

[0145] As described above, the service delivery method adjustment system 1000 of this embodiment predicts virtual data that indicates changes in the subject's situation when a specific service is provided to the subject according to the calculated service delivery method, and then compares a first prediction result obtained by inputting actual data into a nonlinear prediction model with a second prediction result obtained by inputting the virtual data into the nonlinear prediction model, thereby adjusting the calculated service delivery method. This makes it possible to achieve resource allocation, for example, by providing a service to subjects who are likely to benefit from the service provision and halting the service provision to subjects who are not likely to benefit from the service provision. As a result, the resources of the service provider (intervening business operator) can be used more effectively, improving user convenience.

[0146] In other words, the service provision method adjustment system 1000 of this embodiment estimates the intervention method for each intervention recipient based on intervention project information, actual data of the intervention recipient, and prediction results from a non-linear prediction model, generates post-improvement virtual data based on the improvement items due to the intervention and the actual data of the intervention recipient, estimates the effectiveness of the intervention from the difference in the prediction results from the non-linear prediction model, assigns an intervention adjustment flag, and adjusts the intervention method for each intervention recipient, thereby assisting in the selection of the most effective intervention recipient and its intervention method from among the intervention resources available.

[0147] The present invention is not limited to the above-described embodiments and includes various modifications. The above-described embodiments have been described in detail to clearly explain the present invention, and are not necessarily limited to those including all of the described configurations. Furthermore, part of the configuration of one embodiment can be replaced with the configuration of another embodiment. Furthermore, the configuration of another embodiment can be added to the configuration of one embodiment. Furthermore, part of the configuration of each embodiment can be added to, deleted from, or replaced with other configurations.

[0148] The above-described configurations, functions, processing units, processing means, etc. may be partly or entirely implemented in hardware, for example, by designing them as integrated circuits. The above-described configurations, functions, etc. may also be implemented in software, with a processor interpreting and executing a program that implements each function. Information such as the programs, tables, and files that implement each function can be stored in a memory, a hard disk, a recording device such as an SSD (Solid State Drive), an IC card, an SD card, a DVD, or other recording media.

[0149] The service provision method adjustment system 1000 may be configured from a plurality of computers, or may be configured from a single computer.

[0150] The technical features included in the above-described embodiments can be combined as appropriate, not limited to the combinations explicitly stated in the claims.

[0151] For example, in the above embodiment, the configurations expressed as follows are described so that those skilled in the art can implement them.

[0152] (Representation 1) A service provision method adjustment system that adjusts the method of providing a specified service to a target person, which predicts the results of the specified service for the target person based on actual data indicating the target person's status regarding the specified service and a non-linear prediction model, calculates a service provision method that is a method of providing the service to the target person based on the service provision capacity for providing the specified service and the predicted results, predicts virtual data that indicates changes in the target person's status when the specified service is provided to the target person according to the calculated service provision method, and adjusts the calculated service provision method by comparing a first prediction result obtained by inputting the actual data into the non-linear prediction model with a second prediction result obtained by inputting the virtual data into the non-linear prediction model.

[0153] (Representation 2) A service provision method adjustment system according to Representation 1, wherein the real data and the virtual data are data relating to a common specified item among a plurality of items included in the specified service.

[0154] (Representation 3) A service provision method adjustment system according to Representation 1 or 2, which adjusts the calculated service provision method by comparing values of explanatory variables that contribute to the first prediction result among the actual data input to the nonlinear prediction model to predict the first prediction result with values of explanatory variables that contribute to the second prediction result among the virtual data input to the nonlinear prediction model to predict the second prediction result.

[0155] (Representation 4) A service provision method adjustment system described in any one of Representations 1 to 3, in which, when the calculated service provision method is adjusted, an adjustment flag is set indicating that the specified service will be provided to those of the targets to whom the specified service is provided.

[0156] (Representation 5) A service provision method adjustment system described in any one of Representations 1 to 4, in which, when the calculated service provision method is adjusted, an adjustment flag indicating that the specified service will not be provided is set for those of the target individuals to whom the specified service will not be provided.

[0157] (Representation 6) A service provision method adjustment system described in any one of Representations 1 to 5, in which, when the calculated service provision method is adjusted, an adjustment flag is set indicating that the specified service will be provided to those among the targets to whom the specified service will be provided, and an adjustment flag is set indicating that the specified service will not be provided to those among the targets to whom the specified service will not be provided.

[0158] (Representation 7) A service provision method adjustment system according to any one of Representations 1 to 6, wherein the number or percentage of subjects for whom the adjustment flag is set is set based on the service provision capacity.

[0159] (Representation 8) A service providing method adjustment system according to any one of Representations 1 to 7, wherein the nonlinear prediction model is a model that outputs prediction results in an explainable manner.

[0160] (Representation 9) A service provision method adjustment system according to any one of Representations 1 to 8, wherein the specified service is a healthcare service related to the health of the subject.

[0161] (Representation 10) A service provision method adjustment method that adjusts a method of providing a specified service to a target person using a computer, wherein the computer predicts the results of the specified service for the target person based on actual data indicating the target person's status regarding the specified service and a non-linear prediction model, calculates a service provision method that is a method of providing the service to the target person based on the service provision capacity for providing the specified service and the predicted result, predicts virtual data that indicates changes in the target person's status when the specified service is provided to the target person according to the calculated service provision method, and adjusts the calculated service provision method by comparing a first prediction result obtained by inputting the actual data into the non-linear prediction model with a second prediction result obtained by inputting the virtual data into the non-linear prediction model.

[0162] (Representation 11) A service provision method adjustment system that adjusts the method of providing a specified health-related service to a subject, the system generating virtual data after improvement due to the provision of the specified service from actual data regarding the health information of each subject and a table of improvement items due to the provision of the specified service, inputting the actual data regarding the health information of each subject and the virtual data after improvement due to the provision of the specified service into a non-linear prediction model that predicts the subject's risk level, generating prediction result comparison information that compares the prediction result of the actual data with the prediction result of the virtual data based on the prediction results output by the non-linear prediction model, assigning an adjustment flag to each subject from the prediction result comparison information and a standard table for adjusting the provision of the specified service, generating a proposal to adjust the method of providing the specified service for each subject based on a business information table that indicates the service provision capability of a business providing the specified service, the adjustment flag assigned to each subject, and a pre-set adjustment rule, and outputting the generated proposal. [Explanation of symbols]

[0163] 1: Intervention adjustment flag estimation system, 2: Communication network, 3: Intervention history management system, 4: Intervention adjustment rule management system, 5: Intervention adjustment system, 1000: Service provision method adjustment system

Claims

1. A service provision method adjustment system for adjusting a method for providing a predetermined service to a target person, predicting the subject's outcome regarding the predetermined service based on actual data indicating the subject's status regarding the predetermined service and a non-linear prediction model; calculating a service provision method for providing the service to the target person based on the service provision capacity for providing the predetermined service and the predicted result; predicting virtual data indicating a change in the subject's situation when the predetermined service is provided to the subject in accordance with the calculated service provision method; A first prediction result obtained by inputting the actual data into the nonlinear prediction model is compared with a second prediction result obtained by inputting the virtual data into the nonlinear prediction model, and the calculated service provision method is adjusted. Service delivery method coordination system.

2. The real data and the virtual data are data relating to a common predetermined item among a plurality of items included in the predetermined service. The service provision method coordination system according to claim 1 .

3. comparing values of explanatory variables that contribute to the first prediction result among the actual data input to the nonlinear prediction model to predict the first prediction result with values of explanatory variables that contribute to the second prediction result among the virtual data input to the nonlinear prediction model to predict the second prediction result, and adjusting the calculated service provision method; The service provision method adjustment system according to claim 2 .

4. When adjusting the calculated service provision method, an adjustment flag indicating that the predetermined service will be provided to the target persons to whom the predetermined service is to be provided is set. The service provision method adjustment system according to claim 2 .

5. When the calculated service provision method is adjusted, an adjustment flag indicating that the predetermined service will not be provided is set for the target persons to whom the predetermined service will not be provided. The service provision method adjustment system according to claim 2 .

6. When adjusting the calculated service provision method, an adjustment flag is set indicating that the predetermined service will be provided to those of the targets to whom the predetermined service is to be provided, and an adjustment flag is set indicating that the predetermined service will not be provided to those of the targets to whom the predetermined service is not to be provided. The service provision method adjustment system according to claim 2 .

7. The number or percentage of subjects for whom the adjustment flag is set is set based on the service provision capacity. The service provision method adjustment system according to any one of claims 4 to 6.

8. The nonlinear prediction model is a model that outputs predictive results in an explainable manner. The service provision method coordination system according to claim 1 .

9. The predetermined service is a healthcare service related to the health of the subject. The service provision method coordination system according to claim 1 .

10. A service provision method adjusting method for adjusting a method for providing a predetermined service to a target person by a computer, comprising: The computer predicting the subject's outcome regarding the predetermined service based on actual data indicating the subject's status regarding the predetermined service and a non-linear prediction model; calculating a service provision method for providing the service to the target person based on the service provision capacity for providing the predetermined service and the predicted result; predicting virtual data indicating a change in the subject's situation when the predetermined service is provided to the subject in accordance with the calculated service provision method; A first prediction result obtained by inputting the actual data into the nonlinear prediction model is compared with a second prediction result obtained by inputting the virtual data into the nonlinear prediction model, and the calculated service provision method is adjusted. How to adjust service delivery.

11. A service provision method adjustment system for adjusting a method for providing a predetermined health-related service to a subject, generating virtual data after improvement due to the provision of the predetermined service from actual data on health information for each subject and a table of improvement items due to the provision of the predetermined service; Actual data on health information for each subject and virtual data after improvement due to the provision of the predetermined service are input into a nonlinear prediction model that predicts the subject's risk level, generating prediction result comparison information that compares the prediction result of the actual data with the prediction result of the virtual data based on the prediction result output by the nonlinear prediction model; assigning an adjustment flag to each target person based on the prediction result comparison information and a reference table for adjusting the provision of the predetermined service; generating a proposal to adjust the method of providing the predetermined service for each target person based on a business information table indicating the service provision capabilities of a business providing the predetermined service, an adjustment flag assigned to each target person, and a preset adjustment rule; Output the generated proposal Service delivery method coordination system.

Images