Medical checkup result prediction apparatus and medical checkup prediction result presentation apparatus

Abstract

To estimate how a numerical value of a health check item is changed by a future action change. In addition, it is possible to present how the numerical value of the medical examination item changes according to a future behavior change.SOLUTION: The past diagnostic numerical value of the examinee received by a diagnostic numerical value reception part is applied to the health diagnostic result prediction model, and a predicted diagnostic numerical value when the action shown in the action item is present and / or a predicted diagnostic numerical value when the action shown in the action item is absent are estimated for each diagnostic item. In addition, for the diagnosis item and the behavior item selected by the selection unit, a predicted diagnostic value of the diagnosis item when the behavior indicated in the behavior item is present and / or a predicted diagnostic value of the diagnosis item when the behavior indicated in the behavior item is absent is acquired, and the predicted diagnostic value corresponding to the behavior indicated in the behavior item of the diagnosis item acquired by the acquisition unit is presented.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a health checkup result prediction device that predicts health checkup results and a health checkup prediction result presentation device that presents predicted health checkup results. [Background technology]

[0002] Predicting future illnesses and health conditions based on past health checkup results is useful for the health management of examinees. Presenting predicted health checkup results to examinees and their doctors and encouraging them to take appropriate action contributes to preventing illness and maintaining good health. In recent years, health measures for examinees using predicted health checkup results have become extremely important in terms of medical finances and in preventing and treating illness.

[0003] (Prior art 1)

[0004] Prediction methods using past health checkup results include methods for predicting the onset of specific diseases and extracting past items and thresholds that cause health checkup values ​​to worsen. For example, Non-Patent Document 1 describes an invention that determines a reference value that clearly improves the value of maximum blood flow velocity change from the average number of steps taken per day, based on the past health checkup data of 12 people.

[0005] (Prior art 2)

[0006] Another prediction method that uses past health checkup results is a method for predicting increases or decreases in values. For example, Non-Patent Document 2 describes a system that uses a support vector machine to predict whether future health checkup results will "improve," "remain the same," or "worsen" based on a combination of multiple items in past health checkup data. The system uses cross-validation to verify the optimal combination of explanatory variables based on five years of data from 165 people, and explains that it is possible to predict triglyceride levels with an accuracy of approximately 75% to 78%.

[0007] (Prior art 3)

[0008] Furthermore, there is a method for predicting future values ​​themselves as a prediction method using past health checkup results. For example, Non-Patent Document 3 describes an invention that predicts in detail the values ​​of future health checkup items by modeling a relatively complex time series structure with the aim of investigating in detail the correlation between variables related to health checkups.

[0009] (Prior art 4)

[0010] Similar to Conventional Technique 3, Non-Patent Document 4 predicts future numerical values ​​themselves. For example, Non-Patent Document 4 constructs a model that can obtain highly interpretable prediction results. [Prior art documents] [Non-patent literature]

[0011] [Non-Patent Document 1] Nakajima and Takeuchi, "Research on Data Mining Methods in Personalized Healthcare Systems," 10th Forum on Data Engineering and Information Management (DEIM2018), Awara City, Fukui Prefecture, March 2018. [Non-patent document 2] Yamamoto et al., A method for predicting pre-disease status from time-series health checkup data, Information Processing Society of Japan Research Report, 2017-MPS-115 (12), 1-6, published in September 2017. [Non-patent document 3] Hasegawa et al., Future prediction of screening results using a state space model with L1 regularization, 2017 Joint Conference of Statistical Societies, Nanzan University, published in September 2017. [Non-patent document 4] Taniguchi and Tamano, "Construction of a predictive model considering the interpretability of age-related predictions in health checkup predictions," 17th Forum on Information Science and Technology (FIT2018), Fukuoka Institute of Technology, September 2018. Summary of the Invention [Problem to be solved by the invention]

[0012] According to the prior art 1, it is possible to present the average number of steps that should generally be recommended for health promotion, but there is a problem in that it does not take into consideration differences in the characteristics of examinees.

[0013] According to prior art 2, even if it is possible to present predicted results of increases or decreases in the numerical values ​​of future health checkup items, there is a problem in that it is not possible to present the results from the perspective of what actions should be taken to improve the increases or decreases in the numerical values ​​of the items.

[0014] Prior arts 3 and 4 have the problem that although the numerical values ​​of items for future health checkups can be presented, feedback useful for the examinee's health management cannot be obtained.

[0015] All of the prior art techniques either predict the results of future health checkup items if the health checkup items are simply maintained as they are, or at best present standard improvement criteria.

[0016] The present invention aims to make it possible to estimate how the values ​​of health checkup items will change due to future behavioral changes. Another aim of the present invention is to make it possible to present how the values ​​of health checkup items will change in response to future behavioral changes. [Means for solving the problem]

[0017] The first aspect is a health checkup result prediction device comprising: a diagnostic value receiving unit that receives a diagnostic value for each of the examinee's past diagnostic items; a health checkup result prediction model that sets a weight variable indicating the weight to be assigned to changes in the diagnostic value depending on the presence or absence of an action indicated in the action item for each of the diagnostic items depending on the examinee's past diagnostic values, and applies this weight variable to the examinee's past diagnostic values ​​to estimate a predicted diagnostic value that changes depending on the presence or absence of an action for each of the diagnostic items; and an estimation unit that applies the examinee's past diagnostic values ​​received by the diagnostic value receiving unit to the health checkup result prediction model to estimate, for each of the diagnostic items, a predicted diagnostic value when the action indicated in the action item is present and / or a predicted diagnostic value when the action indicated in the action item is absent.

[0018] A second aspect is a health checkup result prediction device in which, in the first aspect, the weight variables are set for each of a plurality of action items for each of a plurality of diagnostic items, the health checkup result prediction model estimates, for each of the plurality of action items for each of the plurality of diagnostic items, a change in a predicted diagnostic numerical value that changes depending on the presence or absence of an action, and the estimation unit estimates, for each of the plurality of action items for each of the plurality of diagnostic items, a predicted diagnostic numerical value when the action indicated in the action item is present and / or a predicted diagnostic numerical value when the action indicated in the action item is absent.

[0019] The third aspect is a health checkup prediction result presentation device comprising: a selection unit that selects any one or more combinations of multiple diagnostic items of a health checkup and any one or more combinations of multiple action items; an acquisition unit that acquires, for the diagnostic items and action items selected by the selection unit, predicted diagnostic values ​​of the diagnostic items when the action indicated in the action item is present and / or predicted diagnostic values ​​of the diagnostic items when the action indicated in the action item is absent; and a presentation unit that presents the predicted diagnostic values ​​acquired by the acquisition unit.

[0020] The fourth aspect is a health checkup prediction result presentation device comprising: a selection unit that selects any one or a combination of two or more of a plurality of diagnostic items of a health checkup; an acquisition unit that selects and acquires, for the diagnostic items selected by the selection unit, a predicted diagnostic value when the behavior indicated in one or more behavior items is present, which is an improved predicted diagnostic value, from among predicted diagnostic values ​​when the behavior indicated in each of a plurality of behavior items is present; and a presentation unit that presents the predicted diagnostic values ​​acquired by the acquisition unit.

[0021] The fifth aspect is a health checkup prediction result presentation device in which, in the fourth aspect, the acquisition unit also acquires predicted diagnostic values ​​when the behavior indicated in the one or more behavior items is absent, and the presentation unit also presents predicted diagnostic values ​​for the diagnostic items when the behavior indicated in the one or more behavior items acquired by the acquisition unit is absent.

[0022] The sixth aspect is a health checkup prediction result presentation device comprising a selection unit that selects any one or a combination of two or more of a plurality of behavioral items of a health checkup, an acquisition unit that selects and acquires predicted diagnostic values ​​for one or more diagnostic items that will result in improved predicted diagnostic values ​​from the predicted diagnostic values ​​for each of a plurality of diagnostic items when the behavior is present for the behavioral item selected by the selection unit, and a presentation unit that presents the predicted diagnostic values ​​acquired by the acquisition unit.

[0023] The seventh aspect is a health checkup prediction result presentation device in which, in the sixth aspect, the acquisition unit also acquires predicted diagnostic values ​​for the one or more diagnostic items when the behavior indicated in the behavior item is absent, and the presentation unit also presents predicted diagnostic values ​​for the one or more diagnostic items acquired by the acquisition unit when the behavior indicated in the behavior item is absent.

[0024] The health checkup prediction result presenting device presents predicted diagnostic values ​​for one or more diagnostic items acquired by the acquisition unit together with predicted diagnostic values ​​for the diagnostic items when the behavior indicated in the behavior item does not occur.

[0025] The eighth aspect is a health checkup prediction result presentation device in any of the third to seventh aspects, which includes a patient identification reception unit that receives the identification of the patient, the acquisition unit acquires the predicted diagnostic values ​​of the specific patient accepted by the patient identification reception unit, and the presentation unit presents the predicted diagnostic values ​​of the specific patient acquired by the acquisition unit.

[0026] A ninth aspect is a health examination prediction result presentation device in any of the third to eighth aspects, wherein the acquisition unit acquires corresponding past diagnostic values ​​along with the predicted diagnostic values, and the presentation unit presents the corresponding past diagnostic values ​​along with the predicted diagnostic values. [Effects of the Invention]

[0027] According to the first and second aspects, it is possible to estimate how the values ​​of the health checkup items will change due to future behavioral changes, and according to the third to ninth aspects, it is possible to present how the values ​​of the health checkup items will change depending on future behavioral changes. [Brief explanation of the drawings]

[0028] [Figure 1] FIG. 1 is a block diagram showing the functional configuration of a medical examination result prediction device according to an embodiment. [Figure 2] FIG. 2 is a block diagram illustrating a functional configuration of the medical checkup prediction result presentation device according to the embodiment. [Figure 3] FIG. 3 is a diagram showing a hardware configuration for realizing the functional configuration of the embodiment shown in FIGS. [Figure 4A] FIG. 4A is a graph showing the estimation results of the estimation unit in association with examinees. [Figure 4B] FIG. 4B is a graph showing the estimation results of the estimation unit in association with examinees. [Figure 4C] FIG. 4C is a graph showing the estimation results of the estimation unit in association with examinees. [Figure 5] FIG. 5 is a graph showing the estimation results of the estimation unit in association with examinees. [Figure 6A] FIG. 6A is a graph showing the estimation results of the estimation unit in association with examinees. [Figure 6B] FIG. 6B is a graph showing the estimation results of the estimation unit in association with examinees. [Figure 6C] FIG. 6A is a graph showing the estimation results of the estimation unit in association with examinees. [Figure 7A] FIG. 7A is a graph showing the estimation results of the estimation unit in association with examinees. [Figure 7B] FIG. 7B is a graph showing the estimation results of the estimation unit in association with examinees. [Figure 8] FIG. 8 is a diagram illustrating a display screen of the display device of the terminal according to the first embodiment. [Figure 9]FIG. 9 is a diagram illustrating a display screen of a display device of a terminal according to the second embodiment. [Figure 10] FIG. 10 is a diagram illustrating a display screen of a display device of a terminal according to the third embodiment. [Figure 11] FIG. 11 is a graph showing the estimation results of the estimation unit in association with examinees. DETAILED DESCRIPTION OF THE INVENTION

[0029] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0030] (Functional configuration of the health examination result prediction device)

[0031] FIG. 1 is a block diagram showing the functional configuration of a medical examination result prediction device 1 according to an embodiment.

[0032] The medical checkup result prediction device of the embodiment includes a diagnostic value receiving unit 100, a medical checkup result prediction model generating unit 200, a medical checkup result prediction model storage unit 250, and an estimation unit 300.

[0033] (Functional configuration of the health checkup prediction result presentation device)

[0034] FIG. 2 is a block diagram showing the functional configuration of the medical checkup prediction result presentation device 2 according to the embodiment.

[0035] The medical checkup prediction result presentation device of the embodiment includes a patient identification reception unit 400, a selection unit 450, an acquisition unit 500, and a presentation unit 600.

[0036] The functions of Figures 1 and 2 can be realized by a combination of one or more personal computer terminals, a server, and a network that connects the terminals and the server so that they can communicate with each other, or by a personal computer terminal alone.

[0037] (Hardware configuration)

[0038] Fig. 3 is a diagram showing a hardware configuration for realizing the functional configuration of the embodiment shown in Fig. 1 and Fig. 2. The hardware configuration of the personal computer terminal 10 or the server is shown as an example.

[0039] As shown in FIG. 3, the personal computer terminal 10 or server includes a CPU (Central Processing Unit) 11, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, storage 14, an input device 16, a display device 17, a communication I / F 18, and an external storage device 19, which are interconnected via a system bus 15 so as to be able to communicate with each other.

[0040] The CPU 11 is a central processing unit that executes various programs and controls each device connected to the system bus 15. That is, the CPU 11 reads a program from the ROM 12 or the storage 14 and executes the program using the RAM 13 as a work area. The CPU 11 controls each device connected to the system bus 15 and performs various arithmetic processing in accordance with the program recorded in the ROM 12 or the storage 14. The ROM 12 or the storage 14 stores a BIOS (Basic Input / Output System) and an OS (Operating System), which are control programs executed by the CPU 11, as well as various computer-readable and executable programs and various necessary data for realizing this embodiment.

[0041] The ROM 12 stores various control programs and various data. The RAM 13 functions as the main memory, work area, etc. of the CPU 11 and temporarily stores programs or data as a working area. The storage 14 is composed of an HDD (Hard Disk Drive) or SSD (Solid State Drive) and stores various programs including the BIOS and OS, and various data.

[0042] The input device 16 includes a pointing device such as a mouse, a keyboard, a reading device such as a scanner, and is used to input various types of information.

[0043] The display device 17 is, for example, a liquid crystal display, and displays various information. The display device 17 may function as the input device 16 by adopting a touch panel system.

[0044] The communication interface 18 is an interface for communicating with other devices such as servers and terminals, and uses standards such as Ethernet (registered trademark), FDDI, Wi-Fi (registered trademark), etc. The communication interface 18 connects to a network and controls the transmission and reception of data.

[0045] The external storage device 19 is configured with various types of memory cards such as USB memory, HDD, SSD, or other external storage media that can be detachably connected.

[0046] (Health checkup result prediction model generation unit)

[0047] The health checkup result prediction model generating unit 200 generates a health checkup result prediction model.

[0048] The procedure for generating a health checkup result prediction model will be described below.

[0049] In the embodiment, a specific and general procedure is shown in which the influence of behaviors that differ for each examinee is extracted from a huge number of diagnostic item data and behavioral item (questionnaire item) data from actual health checkups, and the relationship between behaviors that are effective in promoting health and changes in diagnostic values ​​is calculated.

[0050] In the examples, compared to the conventional technology, a new behavior improvement factor is introduced in addition to the conventional explanatory variables. The behavior improvement factor itself and the interactions between the behavior improvement factor and the conventional explanatory variables are incorporated into the model, and then a variable selection method such as L1 regularization from the conventional technology is combined. In the examples, taking into consideration the addition of a large number of items, namely, the behavior improvement factors, which are categorical data, and their interactions, it is shown that Group Lasso and Sparse Group Lasso are more suitable than Lasso, which is a simple L1 regularization. However, the present invention is not limited to variable selection methods such as L1 regularization.

[0051] In the following, we will use an example of generating a health checkup result prediction model that predicts health checkup results one and two years from the past three years' worth of data based on five years' worth of health checkup data. However, the present invention is not limited to the specific items, number of years of data, or fitting method of the prediction model described below.

[0052] (Collection of health checkup data)

[0053] To generate a health checkup result prediction model, health checkup data shown in 1-1) and 1-2) below will be collected.

[0054] Health checkup data consists of diagnostic numerical data for each diagnostic item and data for each interview item. A health checkup result prediction model is created using health checkup result data from a large number of examinees who form a population over a continuous five-year period. Among the interview items, items that are treated as behavioral change factors are called behavioral items.

[0055] 1-1) Diagnostic numerical data

[0056] The diagnostic numerical data is data on diagnostic values ​​for diagnostic items such as weight, blood pressure, and triglycerides. Five consecutive years of diagnostic numerical data is used. A health checkup result prediction model is constructed that predicts changes in diagnostic values ​​in the fourth and fifth years from the first three years of diagnostic numerical data.

[0057] 1-2) Category item data (medical interview item data)

[0058] The category item data is data from medical examination questionnaires about medication, smoking, drinking, etc. Data from the third year's questionnaire is mainly used to make predictions for the fourth and fifth years. Data from behavioral items, which are part of the questionnaire data, is treated as a factor of behavioral change from the third to fourth year.

[0059] (Pretreatment)

[0060] Preprocessing is a process for improving calculation accuracy, and is not limited to the following method.

[0061] 2-1) nx types of diagnostic numerical data x

[0062] 3 years' worth of diagnostic data for nx types,

[0063]

number

[0064] Here, i indicates the type of diagnostic item (weight, blood pressure, triglyceride, etc.), and j indicates the year.

[0065] If necessary, the diagnostic numerical data xij or the logarithm logxij of the diagnostic numerical value xij is used.

[0066] where:

[0067] yij=logxij

[0068] Let's say.

[0069] 2-2) Calculation of characteristic value z of diagnostic value

[0070] Using three years of diagnostic numerical data xi1, xi2, and xi3 for the same diagnostic item i, perform the following calculation to obtain the characteristic value z of the diagnostic numerical value.

[0071] The average value zi0 is calculated using the following formula:

[0072]

number

[0073] The difference zi1 is calculated using the following formula:

[0074]

number

[0075] The second-order difference zi2 is calculated using the following formula:

[0076]

number

[0077] 2-3) Creating dummy variable data d

[0078] nc type category item data,

[0079]

number

[0080] In contrast,

[0081] ci´=k

[0082] When,

[0083] di´k=1

[0084] otherwise,

[0085] di´k=0

[0086] A dummy variable di'k is created so that: where i' indicates the type of category item (medication, smoking, drinking, etc.), and k indicates the presence of a behavior (the behavior of reducing alcohol consumption) indicated in the category item (e.g., the drinking item). However, to avoid duplication during analysis, the dummy variable with the most frequent value is not used.

[0087] 2-4) Creation of behavior improvement variable data a

[0088] na types of behavioral improvement variable data,

[0089]

number

[0090] Here, i´´ indicates the type of behavioral item (drinking, smoking, taking medicine, etc.).

[0091] First, select behavioral items from the category items and create standards for improving the diagnostic values ​​of the diagnostic items. Items corresponding to behaviors that can be changed at the patient's discretion, such as quitting smoking or reducing alcohol consumption, are selected as behavioral items. For diagnostic value data of diagnostic items such as weight, standards are set for determining improvement, and improvement is determined when a certain level of change is observed.

[0092] For behavioral items, whether or not the diagnostic values ​​of the diagnostic items have improved is determined from the category item data for the third and fourth years, and behavioral improvement variable data ai' indicating whether or not there has been improvement is created.

[0093] The behavior improvement variable data ai'' is set to 1 if there is improvement, and to 0 otherwise (maintaining the current state or worsening of the diagnostic value).

[0094] 2-5) Creating interaction variables m and O

[0095] The interaction variables include a first interaction variable m that indicates the interaction between the behavioral item and the diagnostic value, and a second interaction variable 0 that indicates the interaction between the categorical item and the diagnostic value.

[0096] The first interaction variable is:

[0097]

number

[0098] where,

[0099]

number

[0100] indicates the average value of the characteristic value zij of the diagnostic value, that is, the average value of the characteristic value zij of the diagnostic value of a large number of examinees in the population.

[0101] The first interaction variable mii´´j can be defined as a weight variable that indicates the weight given to changes in diagnostic numerical values ​​depending on the presence or absence of behaviors indicated in the behavior items. This weight variable mii´´j is used as a weight coefficient given to changes in diagnostic numerical values ​​of the examinee in the following formulas (3) and (3)´ that represent the health check result prediction model, as will be described later. In other words, when a behavior indicated in the behavior item is present, the corresponding weight variable mii´´j is used to calculate the examinee's diagnostic numerical value, and when the behavior indicated in the behavior item is absent, the term of the corresponding weight variable mii´´j is set to 0 to calculate the examinee's diagnostic numerical value.

[0102] The weight variable mii´´j is the term on the right side of the above equation (1),

[0103]

number

[0104] As shown in the above, the weight variable mii´´j is set according to the characteristic value zij of the examinee's past diagnostic numerical value. In other words, the greater the deviation between the characteristic value zij of the examinee's past diagnostic numerical value and the average value of the past diagnostic numerical values ​​of a large number of examinees in the population, the larger the weight variable mii´´j becomes. The weight variable mii´´j is set for each diagnostic item i.

[0105] The second interaction variable is:

[0106]

number

[0107] It is expressed as:

[0108] The second interaction variable Oii'jk can be similarly defined as a weight variable indicating the weight given to the change in the diagnostic numerical value depending on the presence or absence of the behavior shown in the behavior item. This weight variable Oii'jk is used as a weight variable giving to the change in the diagnostic numerical value of the examinee in the following formulas (3) and (3)' which show the health check result prediction model, as will be described later. In other words, when the behavior shown in the category item is present, the diagnostic numerical value of the examinee is calculated using the corresponding weight variable Oii'jk, and when the behavior shown in the category item is absent, the term of the corresponding weight variable Oii'jk is set to 0 and the diagnostic numerical value of the examinee is calculated.

[0109] Similarly, the weight variable Oii'jk is set according to the examinee's past diagnosis value zij, and is set for each diagnosis item i.

[0110] (Data Analysis)

[0111] 3-1) Establishment of a health check result prediction model

[0112] Preprocessed data,

[0113]

number

[0114] A health checkup result prediction model is set up to predict the diagnostic value xI4 for each diagnostic item I in the fourth year, which is the objective variable, using as an explanatory variable. Here, θl is a list of all items, including the characteristic value z of the preprocessed diagnostic value, the dummy variable d, the behavior improvement variable a, the first interaction variable m, and the second interaction variable 0, and l is,

[0115]

number

[0116] It is expressed as:

[0117] The mathematical model for predicting health checkup results assumes the following equation (3) or equation (3)', in which both sides of equation (3) are expressed in logarithms.

[0118]

number

[0119]

number

[0120] However, in equations (3) and (3)',

[0121]

number

[0122] is the average value of the patient's diagnostic value for diagnostic item I over the past three years, as shown below.

[0123]

number

[0124] In the above equations (3) and (3)', gl represents the regression coefficient of the explanatory variable θl, h represents the intercept, ε represents the error assuming a standard normal distribution, and σ represents the mean square error.

[0125] 3-2) Calculation of health check result prediction model

[0126] The preprocessed diagnostic numerical value characteristic value z, dummy variable d, behavior improvement variable a, first interaction variable m, and second interaction variable 0 data are applied to equation (3) or equation (3)' to find the unknowns gl, h, σ, and ε. The preprocessed data is used to find the regression coefficient gl, intercept h, and mean error σ that minimize the mean of the squared error σ. However, for combinations of diagnostic numerical data or category item data for the same type of diagnostic item, a Sparse Group Lasso Solver is applied, treating them as the same group, and the Solver parameters can be tuned by cross-validation so that the generalization error, rather than the training error, is minimized.

[0127] In addition, taking into consideration the health promotion effect of presenting future predictive diagnostic results, the following processing will be carried out within the scope that does not lose the validity of the predictive diagnostic results:

[0128] (a) A process of shifting the prediction in the case where the behavior indicated in the behavior item is not performed in a worsening direction in proportion to the expected error.

[0129] (b) A process of exaggerating the improvement effect by a certain amount when the behavior indicated in the behavior item is present.

[0130] It is also effective to carry out the following. In both of the above processes (a) and (b), it is desirable to set values ​​that are likely to occur with a certain probability.

[0131] (Health checkup result prediction model storage section)

[0132] The health checkup result prediction model storage unit 250 stores the health checkup result prediction model generated by the health checkup result prediction model generation unit 200.

[0133] (Diagnostic value reception section)

[0134] The diagnostic value receiving unit 100 receives past diagnostic values ​​for each diagnostic item I of a patient ID1 (ID indicates an identification number, name, etc., and ID1 identifies an individual). In this embodiment, diagnostic values ​​for the past three years for the past diagnostic items I of the patient ID1 are received.

[0135] (Estimation Department)

[0136] The estimation unit 300 applies the past diagnostic numerical values ​​for each diagnostic item I of the examinee ID 1 received by the diagnostic numerical value receiving unit 100 to the health check result prediction model. That is, the estimation unit 300 calculates the explanatory variable θl based on the past diagnostic numerical values ​​for each diagnostic item I of the examinee ID 1 and substitutes it into the formula (3) or (3)', and also calculates the average value of the diagnostic numerical values ​​for the past three years for the diagnostic item I of the examinee ID 1,

[0137]

number

[0138] By substituting this into equation (3) or (3)', the predicted diagnostic value xI4 for each diagnostic item I in the fourth year is calculated. Similarly, the diagnostic value xI5 for each diagnostic item I in the fifth year is calculated based on the diagnostic values ​​for the past three years, including the predicted diagnostic value xI4 for the fourth year.

[0139] This makes it possible to estimate, for each diagnostic item I, the predicted diagnostic value when the behavior indicated in the behavior item is present and the predicted diagnostic value when the behavior indicated in the behavior item is absent.

[0140] For example, let's say diagnostic item I is "triglycerides" and the behavior indicated in the behavior item "drinking" is "reducing alcohol consumption." In this case, the behavior improvement variable is a = 1 (improvement), and the dummy variable is d = 1 (there is behavior to reduce alcohol consumption).

[0141] When the behavior indicated in the behavior item is present, the examinee's diagnostic value is calculated using the weight variables m and O calculated in equations (1) and (2) in equations (3) and (3)', which represent the health check result prediction model.

[0142] On the other hand, when there is no behavior indicated in the behavior item, the weight variables m and O shown in equations (1) and (2) are set to 0 and the examinee's diagnosis value is calculated.

[0143] Fig. 4A is a graph showing the estimation results of estimation unit 300. Fig. 4A shows the estimation results for examinee ID1 identified by ID1. The horizontal axis shows the year, and the vertical axis shows the diagnostic value of a diagnostic item (e.g., triglyceride).

[0144] The diagnostic values ​​of the diagnostic items for the past three years for examinee ID1 are shown by dashed-dotted line LC, the predicted diagnostic values ​​of the diagnostic items for next year (year 4) and the year after next (year 5) when the behavior indicated in the behavior item (e.g., reducing alcohol consumption) is present are shown by solid line LA11, and the predicted diagnostic values ​​of the diagnostic items for next year (year 4) and the year after next (year 5) when the behavior indicated in the behavior item is absent are shown by dashed line LB11. From Figure 4A, it can be seen that the estimation unit 300 estimates, for each diagnostic item (e.g., "triglyceride"), the predicted diagnostic value LA1 when the behavior indicated in the behavior item (e.g., "reducing alcohol consumption") is present and the predicted diagnostic value LB1 when the behavior indicated in the behavior item (e.g., "reducing alcohol consumption") is absent.

[0145] Similarly, Figures 4B and 4C show the estimation results for other examinees ID2 and ID3, respectively, with the same diagnostic items and the same behaviors. Figures 4A to 4C show that the future diagnostic values ​​of each examinee ID1 to ID3 can be estimated with high accuracy based on the past diagnostic values, which differ for each recipient.

[0146] The terminal and server that operate the diagnostic value receiving unit 100, health checkup result prediction model generating unit 200, health checkup result prediction model storage unit 250, and estimation unit 300 can be configured as desired. The diagnostic value receiving unit 100, health checkup result prediction model generating unit 200, health checkup result prediction model storage unit 250, and estimation unit 300 may all be configured to be included in the same terminal, distributed across different terminals, or distributed across different terminals and servers. For example, the health checkup result prediction model storage unit 250 may be provided in a server, and predicted diagnostic values ​​based on the health checkup result prediction model may be obtained by accessing the server from a terminal. Alternatively, the diagnostic value receiving unit 100 may receive diagnostic values ​​for a specific examinee ID1 by accessing a database of examinees provided in the server from a terminal.

[0147] Although an example has been given of estimating a predicted diagnostic value for one diagnostic item ("triglycerides") depending on whether or not a behavior indicated in one behavioral item ("reduced alcohol consumption") is present, it is of course also possible to estimate a predicted diagnostic value for one diagnostic item ("triglycerides") depending on whether or not a behavior indicated in two or more behavioral items ("reduced alcohol consumption," "taking medication," "quitting smoking") is present.

[0148] For example, as shown in Figure 5, it is possible to estimate predicted diagnostic values ​​LA11 and LB11 for the diagnostic item "triglycerides" when the behavior "reducing alcohol consumption" is present and absent, predicted diagnostic values ​​LA12 and LB12 for the diagnostic item "triglycerides" when the behavior "taking medication" is present and absent, and predicted diagnostic values ​​LA13 and LB13 for the diagnostic item "triglycerides" when the behavior "quitting smoking" is present and absent.

[0149] It is also possible to estimate predicted diagnostic values ​​for two or more diagnostic items ("triglycerides," "blood pressure," and "weight") depending on whether or not a behavior indicated in one behavior item ("reduced alcohol consumption") is present.

[0150] For example, as shown in Figure 6A, it is possible to estimate the predicted diagnostic values ​​LA11 and LB11 for the diagnostic item "triglyceride" when the behavior "reducing alcohol consumption" is present and absent, as shown in Figure 6B, it is possible to estimate the predicted diagnostic values ​​LA21 and LB21 for the diagnostic item "blood pressure" when the behavior "reducing alcohol consumption" is present and absent, and further, as shown in Figure 6C, it is possible to estimate the predicted diagnostic values ​​LA31 and LB31 for the diagnostic item "weight" when the behavior "reducing alcohol consumption" is present and absent.

[0151] Similarly, it is also possible to estimate predicted diagnostic values ​​for two or more diagnostic items ("triglycerides," "blood pressure," "weight") depending on whether or not two or more behavioral items ("reducing alcohol consumption," "taking medication," "quitting smoking") are present.

[0152] Alternatively, only the predicted diagnostic value when the behavior indicated in the behavior item is present may be estimated, or only the predicted diagnostic value when the behavior indicated in the behavior item is absent may be estimated.

[0153] For example, as shown in Figure 7A, it is possible to estimate only the predicted diagnostic value LA11 for the diagnostic item "triglyceride" when the behavior "reducing alcohol consumption" is present, the predicted diagnostic value LA12 for the diagnostic item "triglyceride" when the behavior "taking medication" is present, and the predicted diagnostic value LA13 for the diagnostic item "triglyceride" when the behavior "quitting smoking" is present.

[0154] Also, as shown in FIG. 7B, it is possible to estimate only the predicted diagnostic value LB11 for the diagnostic item "neutral fat" when the actions "reducing alcohol consumption," "taking medication," and "quitting smoking" are not taken.

[0155] The estimation results of the estimation unit 300 illustrated in FIGS. 4A to 7B can be displayed on the display screen of a terminal and presented to the examinee or the like.

[0156] Hereinafter, first, second and third embodiments of the health checkup prediction result presenting device will be described.

[0157] Example 1

[0158] 8 is a diagram for explaining Example 1, illustrating a display screen 50 of the display device 17 of the terminal 10. The display screen 50 is configured as an interface screen having the functions of the patient identification reception unit 400, the selection unit 450, and the presentation unit 600 shown in FIG.

[0159] (Patient Specific Reception Department and Selection Department)

[0160] The examinee identification receiving unit 400 receives the identification of the examinee. The selection unit 450 selects any one or a combination of two or more of a plurality of diagnostic items of the health checkup and any one or a combination of two or more of a plurality of behavioral items.

[0161] On the display screen 50, there are arranged a patient identification reception unit 51 corresponding to the patient identification reception unit 400, and a selection unit 52 corresponding to the selection unit 450. The patient identification reception unit 51 receives an ID (identification number, name, etc.) that identifies the patient by a click operation, a text input operation, etc. on the display screen 50. The selection unit 52 selects one or more diagnostic items and one or more behavioral items by a click operation, a text input operation, etc. on the display screen 50. The patient identification reception unit 51 and the selection unit 52 can be configured with a pull-down menu, a check box, a text input box, etc.

[0162] For example, examinee identification reception unit 51 receives examinee ID1, and selection unit 52A selects the diagnostic item "neutral fat," and selection unit 52B selects the behavior item "alcohol drinking."

[0163] (Acquisition Department)

[0164] The acquisition unit 500 acquires predicted diagnostic values ​​for the examinee ID1, whose identification has been accepted by the examinee identification acceptance unit 400 (examinee identification acceptance unit 51), for the diagnostic items and action items selected by the selection unit 450 (selection unit 52). The acquired values ​​are predicted diagnostic values ​​LA11 for the diagnostic items when the action indicated in the action item is present and predicted diagnostic values ​​LB11 for the diagnostic items when the action indicated in the action item is absent. The acquisition method may be either calculation processing or data reading. For example, the calculated predicted diagnostic values ​​can be acquired by applying the past diagnostic values ​​of the examinee ID1 to the health check result prediction model shown in equation (3) or (3)'. Alternatively, for example, data on predicted diagnostic values ​​for each examinee ID1, ID2, etc., calculated using the health check result prediction model shown in equation (3) or (3)' may be stored in a database in advance, and the predicted diagnostic values ​​associated with the examinee ID1 can be acquired from the database.

[0165] The acquisition unit 500 also acquires the past diagnosis values ​​LC of the diagnosis items selected by the selection unit 450 (selection unit 52) ​​of the specific examinee ID1.

[0166] (Presentation part)

[0167] The presentation unit 600 presents the predicted diagnostic values ​​LA11 and LB11 corresponding to the behavior of "reducing alcohol consumption" for the diagnostic item of "triglyceride" acquired by the acquisition unit 500. The presentation unit 600 can present the corresponding past diagnostic values ​​LC together with the predicted diagnostic values ​​LA11 and LB11.

[0168] The predicted diagnostic values ​​LA11 and LB11 are displayed together with the past diagnostic value LC on the display unit 53 of the display screen 50 of the display device 17. For example, as shown in Fig. 4A, the display unit 53 displays the predicted diagnostic values ​​LA11 and LB11 for the diagnostic item "triglyceride" for examinee ID1 when the behavior "reducing alcohol intake" is present and absent, together with the past diagnostic value LC.

[0169] Furthermore, when the diagnostic item "triglyceride" is selected in the selection unit 52 and the behavior items indicating the behaviors "reduce alcohol intake," "take medication," and "quit smoking" are selected, the display unit 53 displays the display shown in Fig. 5. That is, as shown in Fig. 5, the predicted diagnostic values ​​LA11 and LB11 of the diagnostic item "triglyceride" when the behavior "reduce alcohol intake" is present and absent are displayed together with the past diagnostic value LC, the predicted diagnostic values ​​LA12 and LB12 of the diagnostic item "triglyceride" when the behavior "take medication" is present and absent are displayed together with the past diagnostic value LC, and the predicted diagnostic values ​​LA13 and LB13 of the diagnostic item "triglyceride" when the behavior "quit smoking" is present and absent are displayed together with the past diagnostic value LC.

[0170] Furthermore, when the diagnostic items "triglyceride," "blood pressure," and "body weight" are selected in the selection unit 52, and an action item indicating the action of "reducing alcohol consumption" is selected, the displays shown in Figures 6A, 6B, and 6C are displayed on the display unit 53. That is, as shown in Figure 6A, predicted diagnostic values ​​LA11 and LB11 for the diagnostic item "triglyceride" when the action of "reducing alcohol consumption" is present and absent are displayed together with past diagnostic values ​​LC, as shown in Figure 6B, predicted diagnostic values ​​LA21 and LB21 for the diagnostic item "blood pressure" when the action of "reducing alcohol consumption" is present and absent are displayed together with past diagnostic values ​​LC, and as shown in Figure 6C, predicted diagnostic values ​​LA31 and LB31 for the diagnostic item "body weight" when the action of "reducing alcohol consumption" is present and absent are displayed together with past diagnostic values ​​LC.

[0171] Similarly, when the diagnostic items "triglycerides," "blood pressure," and "weight" are selected in the selection unit 52, and the behavioral items indicating the actions of "reducing alcohol intake," "taking medication," and "quitting smoking" are selected, the corresponding display is made on the display unit 53.

[0172] In addition, when a diagnostic item such as "triglyceride" is selected in the selection unit 52 and an action item indicating an action such as "reducing alcohol consumption," "taking medication," or "quitting smoking" is selected, only the presence of the action may be displayed on the display unit 53. For example, as shown in Fig. 7A, the predicted diagnostic value LA11 of the diagnostic item "triglyceride" when the action "reducing alcohol consumption" is present can be displayed together with the past diagnostic value LC, the predicted diagnostic value LA12 of the diagnostic item "triglyceride" when the action "taking medication" is present can be displayed together with the past diagnostic value LC, and the predicted diagnostic value LA13 of the diagnostic item "triglyceride" when the action "quitting smoking" is present can be displayed together with the past diagnostic value LC.

[0173] Only when no behavior has been performed may be displayed on the display unit 53. For example, as shown in Fig. 7B, the predicted diagnostic value LB11 for the diagnostic item "neutral fat" when the behaviors "reducing alcohol intake," "taking medication," and "quitting smoking" have not been performed can be displayed together with the past diagnostic value LC.

[0174] Example 2

[0175] 9 is a diagram for explaining the second embodiment, illustrating an example of a display screen 50 of the display device 17 of the terminal 10. In the following, the same components as those in the first embodiment are given the same reference numerals, and the description thereof will be omitted as appropriate.

[0176] (Patient Specific Reception Department and Selection Department)

[0177] The examinee identification receiving unit 400 receives the identification of the examinee. The selection unit 450 selects one or a combination of two or more of the multiple diagnostic items for the health checkup.

[0178] For example, the examinee identification reception unit 51 receives an examinee with ID1, and the selection unit 52 selects the diagnosis item "blood pressure."

[0179] (Acquisition Department)

[0180] The acquiring unit 500 selects and acquires predicted diagnostic values ​​for the examinee ID1, whose identification has been accepted by the examinee identification accepting unit 400 (examinee identification accepting unit 51), when there are actions indicated in one or more action items that result in improved predicted diagnostic values, among predicted diagnostic values ​​when there are actions indicated in each of the multiple action items for the diagnostic item selected by the selecting unit 450 (selecting unit 52). For example, the acquiring unit 500 selects and acquires predicted diagnostic values ​​LA22, LA21, and LA23 when there are the top three actions "weight loss," "reduction in alcohol intake," and "reduction in waist circumference," which result in improved predicted diagnostic values ​​for the diagnostic item "blood pressure."

[0181] The acquisition unit 500 also acquires the past diagnosis values ​​LC of the diagnosis items selected by the selection unit 450 (selection unit 52) ​​of the specific examinee ID1.

[0182] (Presentation part)

[0183] The presentation unit 600 presents predicted diagnostic values ​​LA22, LA21, and LA23 when the top three actions "weight loss," "reduced alcohol intake," and "reduced waist circumference" are present, which improve the predicted diagnostic value of the diagnostic item "blood pressure" acquired by the acquisition unit 500. The presentation unit 600 can present the corresponding past diagnostic values ​​LC together with the predicted diagnostic values ​​LA22, LA21, and LA23.

[0184] The predicted diagnostic values ​​LA22, LA21, and LA23 are displayed together with past diagnostic values ​​and LC on the display unit 53 of the display screen 50 of the display device 17. For example, as shown in Fig. 11, the display unit 53 displays the predicted diagnostic values ​​LA22 and LB22 of the diagnostic item "blood pressure" when the behavior "weight loss" is present and absent, along with the past diagnostic value LC, the predicted diagnostic values ​​LA21 and LB21 of the diagnostic item "blood pressure" when the behavior "reduce alcohol intake" is present and absent, along with the past diagnostic value LC, and the predicted diagnostic values ​​LA23 and LB23 of the diagnostic item "blood pressure" when the behavior "reduce waist circumference" is present and absent, along with the past diagnostic value LC.

[0185] Note that only when an action has been taken may be displayed on the display unit 53. That is, the display unit 53 can display the predicted diagnostic value LA22 of the diagnostic item "blood pressure" when the action "weight loss" has been taken together with the past diagnostic value LC, the predicted diagnostic value LA21 of the diagnostic item "blood pressure" when the action "reduce alcohol consumption" has been taken together with the past diagnostic value LC, and the predicted diagnostic value LB23A of the diagnostic item "blood pressure" when the action "reduce waist circumference" has been taken together with the past diagnostic value LC.

[0186] The display unit 53 may display only when no behavior is taking place. That is, the display unit 53 can display the predicted diagnostic values ​​LB22, LB21, and LB23 for the diagnostic item "blood pressure" when the behaviors "weight loss," "reduced alcohol intake," and "reduced waist circumference" are not taking place, together with the past diagnostic value LC.

[0187] Example 3

[0188] FIG. 10 is a diagram for explaining the third embodiment, illustrating an example of a display screen 50 of the display device 17 of the terminal 10. In FIG.

[0189] (Patient Specific Reception Department and Selection Department)

[0190] The examinee identification receiving unit 400 receives the identification of the examinee. The selection unit 450 selects one or a combination of two or more of the multiple diagnostic items for the health checkup.

[0191] For example, the examinee identification reception unit 51 receives an examinee with ID1, and the selection unit 52 selects the behavior item "drinking alcohol."

[0192] (Acquisition Department)

[0193] The acquisition unit 500 selects and acquires predicted diagnostic values ​​for one or more diagnostic items that are improved predicted diagnostic values ​​from among the predicted diagnostic values ​​for each of the multiple diagnostic items when the behavior is present, for the behavior item selected by the selection unit 450 (selection unit 52) ​​of the examinee ID1 whose identification has been accepted by the examinee identification acceptance unit 400 (examinee identification acceptance unit 51). For example, the acquisition unit 500 selects and acquires predicted diagnostic values ​​LA11, LA21, and LA31 for the top three diagnostic items, "triglycerides," "blood pressure," and "weight," that are improved predicted diagnostic values ​​when the behavior "reducing alcohol intake" is present.

[0194] The acquisition unit 500 also acquires the past diagnosis values ​​LC of the diagnosis items selected by the selection unit 450 (selection unit 52) ​​of the specific examinee ID1.

[0195] (Presentation part)

[0196] The presentation unit 600 presents predicted diagnostic values ​​LA11, LA21, and LA31 of the top three diagnostic items, "triglycerides," "blood pressure," and "weight," whose predicted diagnostic values ​​are improved when the behavior of "reducing alcohol consumption" acquired by the acquisition unit 500 is present. The presentation unit 600 can present the corresponding past diagnostic values ​​LC together with the predicted diagnostic values.

[0197] The predicted diagnostic values ​​are displayed together with past diagnostic values ​​on the display unit 53 of the display screen 50 of the display device 17. For example, as shown in Fig. 6A, the display unit 53 displays predicted diagnostic values ​​LA11 and LB11 for the diagnostic item "triglyceride" when the behavior "reducing alcohol intake" is present and absent, along with the past diagnostic value LC; as shown in Fig. 6B, the display unit 53 displays predicted diagnostic values ​​LA21 and LB21 for the diagnostic item "blood pressure" when the behavior "reducing alcohol intake" is present and absent, along with the past diagnostic value LC; and as shown in Fig. 6C, the display unit 53 displays predicted diagnostic values ​​LA31 and LB31 for the diagnostic item "weight" when the behavior "reducing alcohol intake" is present and absent, along with the past diagnostic value LC.

[0198] Note that the display unit 53 may only display when the behavior is present. That is, the display unit 53 can display the predicted diagnostic value LA11 for the diagnostic item "triglyceride" when the behavior "reducing alcohol consumption" is present together with the past diagnostic values ​​LC, the predicted diagnostic value LA21 for the diagnostic item "blood pressure" when the behavior "reducing alcohol consumption" is present together with the past diagnostic values ​​LC, and the predicted diagnostic value LA31 for the diagnostic item "weight" when the behavior "reducing alcohol consumption" is present together with the past diagnostic values ​​LC.

[0199] Only when no behavior is performed may be displayed on the display unit 53. That is, the display unit 53 can display the predicted diagnostic values ​​LB11, LB21, and LB31 of the diagnostic items "neutral fat," "blood pressure," and "weight" when the behavior "reducing alcohol consumption" is not performed, together with the past diagnostic values ​​LC.

[0200] (Effects of the embodiment)

[0201] According to the embodiment, it is possible to estimate how the values ​​of health checkup items will change due to future behavioral changes. Also, it is possible to present how the values ​​of health checkup items will change depending on future behavioral changes. Therefore, the effect of encouraging the user to take action to improve the values ​​is enhanced compared to conventional techniques. [Explanation of symbols]

[0202] 100 Diagnostic Value Reception Section 200 Health check result prediction model generation unit 250 Health check result prediction model storage section 300 Estimation section 400 Selection Section 500 Acquisition Department 600 Presentation section

Claims

1. a selection unit for selecting one or a combination of two or more of a plurality of diagnostic items of the health checkup and one or a combination of two or more of a plurality of behavioral items; an acquisition unit that acquires, for the diagnostic items and action items selected by the selection unit, predicted diagnostic values ​​of the diagnostic items when the action indicated in the action item is present and / or predicted diagnostic values ​​of the diagnostic items when the action indicated in the action item is absent; a presentation unit that presents the predicted diagnostic values ​​acquired by the acquisition unit; A health checkup prediction result presentation device comprising:

2. a selection unit for selecting one or a combination of two or more of a plurality of diagnostic items of the health checkup; an acquisition unit that selects and acquires a predicted diagnostic value when the behavior indicated in one or more action items is present, which is an improved predicted diagnostic value, from predicted diagnostic values ​​when the behavior indicated in each of a plurality of action items is present, for the diagnostic item selected by the selection unit; a presentation unit that presents the predicted diagnostic values ​​acquired by the acquisition unit; A health checkup prediction result presentation device comprising:

3. The acquisition unit and acquiring a predicted diagnostic value when the behavior indicated in the one or more behavior items is not performed; The presentation unit and presenting a predicted diagnostic value of the diagnostic item when the behavior indicated in the one or more behavior items acquired by the acquisition unit is absent. The health checkup prediction result presentation device according to claim 2 .

4. a selection unit for selecting one or a combination of two or more of a plurality of behavioral items for the health checkup; an acquisition unit that selects and acquires predicted diagnostic values ​​for one or more diagnostic items that will be improved predicted diagnostic values ​​from predicted diagnostic values ​​for each of a plurality of diagnostic items when the behavior is present, for the behavior item selected by the selection unit; a presentation unit that presents the predicted diagnostic values ​​acquired by the acquisition unit; A health checkup prediction result presentation device comprising:

5. The acquisition unit and acquiring a predicted diagnostic value for the absence of the behavior indicated in the behavior item for the one or more diagnostic items; The presentation unit and presenting predicted diagnostic values ​​of the diagnostic items when the behavior indicated in the behavior item does not exist for one or more diagnostic items acquired by the acquisition unit. The health checkup prediction result presentation device according to claim 4 .

6. a patient identification reception unit that receives identification of a patient; the acquiring unit acquires predicted diagnostic values ​​of the specific examinee accepted by the examinee identification accepting unit, The presentation unit presents the predicted diagnostic values ​​of the specific examinee acquired by the acquisition unit. The health checkup prediction result presentation device according to claim 1 .

7. The acquisition unit acquires the corresponding past diagnostic values ​​together with the predicted diagnostic values, The presentation unit presents the corresponding past diagnostic values ​​together with the predicted diagnostic values. The health checkup prediction result presentation device according to claim 1 .

Images