Estimation system, method for estimation, and estimation program

Abstract

To provide an estimation system, a method for estimation, and an estimation program which can improve the accuracy of estimating the likelihood of falling of a user.SOLUTION: An estimation system 100 for estimating the user's likelihood of falling includes an acquisition unit, a storage unit, and a generation unit. The acquisition unit acquires evaluation information including oral information regarding the oral function of the user. The storage unit stores a classification model derived by using plural pieces of reference evaluation information and a fall history linked to the reference evaluation information. The generation unit refers to the classification model and generates fall information indicating the user's possibility of falling based on the evaluation information.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to an estimation system, an estimation method, and an estimation program. [Background technology]

[0002] Conventionally, as a method for estimating the possibility that a user will fall, a data analysis system such as that disclosed in Patent Document 1 has been proposed.

[0003] Patent Document 1 discloses a data analysis system that sets multiple evaluation items related to a specified case, evaluates learning data in which classification information related to the relevance of each of the multiple evaluation items is set, evaluates target data for each of the multiple evaluation items based on the evaluation results of the learning data, performs analysis based on a specified prediction model using the evaluation of the target data for each of the multiple evaluation items as input, and outputs predictive information regarding the occurrence of a specified case based on the results of the analysis. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] International Publication No. 2017 / 199445 Summary of the Invention [Problem to be solved by the invention]

[0005] Here, the technology disclosed in Patent Document 1 is premised on predicting a patient's fall based on the target data, which is the content recorded in an electronic medical record or the like. However, the content recorded in an electronic medical record or the like may contain a lot of content whose relevance to falls is unclear, or content written in different expressions for each electronic medical record. For this reason, there is a concern that the publicly known technology such as the technology disclosed in Patent Document 1 may have poor accuracy in estimating the possibility of a fall.

[0006] The present invention has been devised in consideration of the above-mentioned problems, and its object is to provide an estimation system, an estimation method, and an estimation program that improve the accuracy of estimating the possibility of a user falling. [Means for solving the problem]

[0007] The estimation system of the first invention is an estimation system that estimates the possibility of a user falling, and is characterized in that it comprises an acquisition unit that acquires evaluation information including oral information related to the user's oral function, a storage unit that stores a classification model derived using a plurality of reference evaluation information and a fall history linked to the reference evaluation information, and a generation unit that refers to the classification model and generates fall information indicating the user's possibility of falling in relation to the evaluation information.

[0008] The estimation system according to a second aspect of the present invention is characterized in that, in the first aspect of the present invention, the evaluation information includes subjective information related to the user's subjective health state.

[0009] The estimation system according to a third aspect of the present invention is characterized in that, in the second aspect, the evaluation information includes physical information related to the user's physical function and cognitive information related to the user's cognitive function.

[0010] The estimation system according to a fourth invention is characterized in that, in the third invention, the classification model represents a logistic regression model derived with the reference evaluation information as an explanatory variable and the fall history as a dependent variable, and the generation unit generates the fall information using the results of assigning odds indicating the degree of contribution of the oral information, the subjective information, the physical information, and the cognitive information contained in the evaluation information to estimating the possibility of a fall.

[0011] The estimation system according to the fifth invention is any one of the first to fourth inventions, characterized in that the storage unit stores a plurality of pieces of suggested information regarding fall prevention measures linked to the possibility of falling, and the generation unit selects the suggested information based on the generated fall information.

[0012] The estimation method of the sixth invention is an estimation method for estimating the possibility of a user falling, and is characterized in that it includes an acquisition step of acquiring evaluation information including oral information related to the user's oral function, and a generation step of referring to a classification model derived using a plurality of reference evaluation information and a fall history linked to the reference evaluation information, and generating fall information indicating the possibility of the user falling in relation to the evaluation information.

[0013] The estimation program of the seventh invention is an estimation program that estimates the possibility of a user falling, and is characterized in that it causes a computer to execute an acquisition step of acquiring evaluation information including oral information related to the user's oral function, and a generation step of referring to a classification model derived using a plurality of reference evaluation information and a fall history linked to the reference evaluation information, and generating fall information indicating the possibility of the user falling in relation to the evaluation information. Effect of the Invention

[0014] According to the first to fifth inventions, the acquisition unit acquires evaluation information including oral cavity information related to the oral cavity function of the user. The generation unit refers to the classification model and generates fall information indicating the possibility of the user falling in response to the evaluation information. Therefore, the possibility of the user falling can be estimated based on the oral cavity information that is likely to affect falls. This makes it possible to improve the accuracy of estimating the possibility of the user falling.

[0015] In particular, according to the second aspect of the present invention, the evaluation information includes subjective information on the user's subjective health condition. Therefore, the possibility of the user falling can be estimated based on a combination of oral cavity information, which is likely to affect a fall, and the subjective information. This makes it possible to further improve the accuracy of estimating the possibility of the user falling.

[0016] In particular, according to the third aspect of the present invention, the evaluation information includes physical information related to the user's physical functions and cognitive information related to the user's cognitive functions. Therefore, the possibility of the user falling can be estimated from multiple perspectives based on the four types of information that are likely to affect falls. This makes it possible to further improve the accuracy of estimating the possibility of the user falling.

[0017] In particular, according to the fourth aspect of the present invention, the generation unit generates fall information using the results of assigning odds to the oral cavity information, subjective information, physical information, and cognitive information included in the evaluation information. Therefore, the user's likelihood of falling can be estimated based on the degree to which each piece of information contributes to the estimation of the likelihood of falling. This makes it possible to further improve the accuracy of estimating the user's likelihood of falling.

[0018] In particular, according to the fifth aspect of the present invention, the generation unit selects the proposed information based on the generated fall information. This can lead to a reduction in the possibility of the user falling. This makes it possible to improve the maintenance of the user's health.

[0019] According to the sixth and seventh aspects of the invention, the acquisition step acquires evaluation information including oral cavity information related to the oral cavity function of the user. The generation step refers to the classification model and generates fall information indicating the possibility of the user falling for the evaluation information. Therefore, the possibility of the user falling can be estimated based on the oral cavity information that is likely to affect falls. This makes it possible to improve the accuracy of estimating the possibility of the user falling. [Brief description of the drawings]

[0020] [Figure 1] FIG. 1 is a schematic diagram illustrating an example of an estimation system according to an embodiment. [Diagram 2] FIG. 2 is a schematic diagram showing an example of a checklist used when acquiring evaluation information. [Diagram 3] FIG. 3(a) is a schematic diagram showing an example of the configuration of an estimation device, and FIG. 3(b) is a schematic diagram showing an example of the function of the estimation device. [Figure 4] FIG. 4 is a flowchart illustrating an example of an estimation method according to an embodiment. [Diagram 5] Figure 5 shows the 12 items evaluated for their relationship with fall outcomes. [Figure 6] FIG. 6 shows eight items evaluated for their relationship with fall outcomes. [Figure 7] FIG. 7 shows eight items evaluated for their relationship with fall outcomes. [Figure 8] FIG. 8 shows the nine items evaluated for their relationship with fall outcomes. [Figure 9] Figure 9 shows the 23 items evaluated for their tendency to fall probability using the classification model. [Figure 10] Figure 10 shows the 21 items evaluated for their tendency to fall probability using the classification model. [Figure 11] FIG. 11 is a diagram showing the items and features used when creating the two types of models. [Figure 12] FIG. 12 is a diagram showing the evaluation results of Model 1. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0021] Hereinafter, an example of an estimation system, an estimation method, and an estimation program according to an embodiment of the present invention will be described with reference to the drawings.

[0022] (Embodiment: Estimation system 100) FIG. 1 is a schematic diagram showing an example of an estimation system 100 according to the present embodiment.

[0023] The estimation system 100 is used to estimate the possibility that a user will fall. By using the estimation system 100, measures to prevent the user from falling can be considered, and the estimation system 100 can also lead to improved motivation to perform exercises that are effective in reducing the possibility of falling.

[0024] The estimation system 100 includes an estimation device 1, as shown in Fig. 1, and may also include, for example, a user terminal 2 and a server 4 connected to the estimation device 1. The estimation system 100 acquires, as evaluation information, responses to questions written in a checklist (e.g., Fig. 2) or questionnaire targeted at the user, and measurement results. Then, the estimation system 100 refers to a classification model derived in advance and generates fall information indicating the possibility of the user falling based on the acquired evaluation information.

[0025] <Evaluation information> The evaluation information includes oral cavity information and may include at least one of subjective information, physical information, cognitive information, and basic information. The evaluation information may be obtained based on responses to a questionnaire or the like targeted at users, or may be obtained from a known measuring device such as a weight scale.

[0026] The evaluation information may be, for example, a binary value based on a two-level evaluation such as "yes" or "no" obtained as an answer to a question item, or a multi-level evaluation such as "6: very bad", "5: bad", "4: not very good", "3: average", "2: somewhat good", or "1: very good", and may be set in any combination according to the contents of the questionnaire. Note that, for example, items from the Basic Survey on Living Conditions of the People or information already compiled by a specific group or organization may be used as the evaluation information.

[0027] <<Oral information>> The oral cavity information indicates information related to the oral cavity function of the user. The oral cavity information indicates a known degree (state) of oral cavity function, and includes, for example, at least one of a degree of eating, a degree of chewing, and a degree of swallowing. The oral cavity information may include, for example, at least one of a degree of saliva secretion, a degree of taste, a degree of articulation function, and a number of teeth.

[0028] The inventors have newly discovered that the possibility of falling increases with the decline in oral cavity function. Therefore, the estimation system 100 can improve the accuracy of estimating the possibility of a user falling by using evaluation information including oral cavity information.

[0029] In particular, by including the degree of swallowing as oral cavity information, it is possible to easily improve the accuracy of estimating the possibility of a user falling. For example, an answer to a question such as "Do you ever choke on tea, soup, etc." may be used as the degree of swallowing. In addition to the above, answers to questions for grasping the tendency of swallowing disorders, such as "I often choke while eating," "It has become difficult to eat hard foods," "I have been clearing my throat more often," and "I feel like food is stuck in my throat," may also be used as the degree of swallowing.

[0030] <<Subjective information>> The subjective information indicates information related to the user's subjective health. The subjective information includes the user's own perception of his or her daily health condition and the degree of regular exercise, etc. The subjective information includes answers to questions such as "How has your health been over the past year?" and "How many years have you been doing a particular type of exercise?"

[0031] <<Physical information>> The physical information indicates information related to the user's physical functions. The physical information includes at least one of the measurement results and the evaluation results of the user's physical abilities. For example, the physical information includes, as a measurement result, "time spent standing on one leg with eyes open," and as an evaluation result, answers to questions such as "Can you stand up from a chair without using your hands?" and "Have you fallen in the past year?"

[0032] <<Cognitive information>> The cognitive information indicates information about the cognitive function of the user. The cognitive information includes the degree of the user's intellectual function, such as the user's ability to understand and judge. The cognitive information includes, for example, answers to questions such as "Do you sometimes not know what date it is today?"

[0033] <<Basic information>> The basic information indicates information about the basic characteristics of the user. The basic information includes information indicating personal characteristics such as the user's age, sex, weight, etc., and information about the living conditions. The basic information includes, for example, answers to questions such as "Are you receiving care?"

[0034] The basic information has a smaller effect on the estimation of the likelihood of a fall than the other pieces of information described above. On the other hand, classification for each specific user based on the basic information can lead to improved accuracy in estimating the likelihood of a fall.

[0035] <Classification model> The classification model is derived using a plurality of pieces of reference evaluation information and a fall history linked to the reference evaluation information. The classification model indicates a function (regression model) obtained by regression analysis. The classification model indicates, for example, a logistic regression model derived with the reference evaluation information as an explanatory variable and the fall history as an objective variable. The classification model may be a trained model derived using machine learning such as deep learning.

[0036] The reference evaluation information indicates evaluation information acquired in advance. The fall history indicates whether the user who is the subject of the reference evaluation information has fallen before or after the reference evaluation information is created. As the fall history, the number of falls in one year after the reference evaluation information is created may be used, or physical information included in the reference evaluation information (e.g., an answer to the question "Have you fallen in the past year?") may be used.

[0037] In the estimation system 100, for example, a plurality of classification models may be used. The plurality of classification models differ, for example, in the derivation method and the reference evaluation information used during derivation. The plurality of classification models are derived, for example, using the reference evaluation information classified based on the basic information and the fall history. In this case, for example, a classification model suitable for each age and sex of the user can be derived.

[0038] (Estimation device 1) FIG. 3(a) is a schematic diagram showing an example of the configuration of the estimation device 1, and FIG. 3(b) is a schematic diagram showing an example of the function of the estimation device 1.

[0039] A known electronic device such as a personal computer, a smartphone, or a tablet terminal is used as the estimation device 1. As shown in FIG. 3(a), the estimation device 1 includes a housing 10, a CPU (Central Processing Unit) 101, a ROM (Read Only Memory) 102, a RAM (Random Access Memory) 103, a storage unit 104, and I / Fs 105 to 107. The components 101 to 107 are connected to each other via an internal bus 110.

[0040] The CPU 101 controls the entire estimation device 1. The ROM 102 stores operation code for the CPU 101. The RAM 103 is a working area used when the CPU 101 is operating.

[0041] Various types of information such as evaluation information and classification models are stored in the storage unit 104. As the storage unit 104, for example, a data storage device such as an SSD (Solid State Drive) is used.

[0042] The I / F 105 is an interface for transmitting and receiving various information to and from the user terminal 2, the server 4, etc., as necessary, via the communication network 3. The I / F 106 is an interface for transmitting and receiving information to and from the input unit 108. For example, a keyboard or the like is used as the input unit 108. The I / F 107 is an interface for transmitting and receiving various information to and from the display unit 109. A display is used as the display unit 109, and in the case of a touch panel type, for example, it is provided integrally with the input unit 108.

[0043] 3(b), the estimation device 1 includes an acquisition unit 11, a generation unit 12, and an output unit 13, and may include, for example, a storage unit 14. Each function shown in FIG. 3(b) is realized by the CPU 101 using the RAM 103 as a working area to execute a program stored in the storage unit 104 or the like.

[0044] <Acquisition part 11> The acquisition unit 11 acquires evaluation information including oral cavity information. The acquisition unit 11 acquires evaluation information input by a user or the like via the input unit 108, and may also acquire evaluation information transmitted from, for example, the user terminal 2 or the server 4.

[0045] <Generation part 12> The generating unit 12 generates fall information for the evaluation information by referring to the classification model. The generating unit 12 may generate fall information for the evaluation information by referring to a classification model that differs for each piece of information included in the evaluation information, for example.

[0046] For example, the fall information is generated in a data format that can be recognized by the user, and can be set to any format depending on the purpose. As the fall information, for example, a character string indicating the estimated possibility of falling (e.g., a percentage) and the estimated period, such as "XX% possibility of falling within one year", is used. In particular, by making the estimated period equal to the target period in various information included in the evaluation information, it is possible to improve the estimation accuracy.

[0047] The generation unit 12 may select the suggested information based on the generated fall information, for example. The suggested information indicates content for reducing the possibility of the user falling, such as "By doing □□ every morning, the possibility of falling tends to be reduced."

[0048] The proposed information indicates information on fall prevention measures linked to the possibility of falling. A plurality of pieces of proposed information are stored in the storage unit 104, for example, and proposed information suitable for the content of the fall information is selected. For example, a pair of a fall history and reference proposed information linked to the fall history may be stored in the storage unit 104, and the fall history corresponding to the fall information generated by the generation unit 12 may be identified, and the reference proposed information linked to the identified fall history may be selected as the proposed information.

[0049] The suggested information may include, for example, information for improving the degree of swallowing. The suggested information may include, for example, a method for preventing aspiration pneumonia (for example, a method for improving muscle strength of the tongue or throat).

[0050] <Output section 13> The output unit 13 outputs the generated fall information. The output unit 13 outputs the fall information to the display unit 109, and may also output the fall information to the user terminal 2, the server 4, or the like, for example.

[0051] <Storage section 14> The storage unit 14 retrieves, as necessary, various pieces of information stored in the storage unit 104. The storage unit 14 stores, as necessary, various pieces of information used by each of the components 11 to 13 in the storage unit 104.

[0052] (User terminal 2) The user terminal 2 refers to a terminal owned by a user. As the user terminal 2, for example, a known electronic device such as a personal computer, a smartphone, or a tablet terminal is used.

[0053] For example, an application for generating evaluation information may be installed in the user terminal 2. The application may include a function for allowing the user to answer questions such as those shown in FIG.

[0054] The user terminal 2 may, for example, have at least some of the functions of the above-mentioned estimation device 1. That is, the user terminal 2 may execute, for example, processes from obtaining evaluation information to generating fall information.

[0055] (Communication Network 3) The communication network 3 is an internet network or the like for connecting the user terminal 2, the server 4, etc. to the estimation device 1. The communication network 3 indicates known wireless communication or wired communication, and can be arbitrarily used depending on the application.

[0056] (Server 4) The server 4 saves and accumulates various types of information, such as evaluation information and fall information, transmitted via the communication network 3. Based on a request from the estimation device 1, the server 4 may transmit the accumulated information to the estimation device 1 via the communication network 3. The server 4 may be connected to, for example, a plurality of estimation devices 1 or a plurality of user terminals 2, and may acquire various types of information from each estimation device 1 or each user terminal 2 and store it collectively.

[0057] The server 4 may have at least some of the functions of the above-mentioned estimation device 1. That is, the server 4 may execute processes from obtaining evaluation information to generating fall information, for example.

[0058] (Embodiment: Estimation method) Next, an example of the estimation method in this embodiment will be described with reference to a flowchart shown in FIG.

[0059] The estimation method includes an acquisition step S110 and a generation step S120. At least a part of the estimation method is implemented, for example, by using the estimation system 100, and can be executed, for example, via an estimation program installed in the estimation device 1, the user terminal 2, or the server 4.

[0060] <Acquisition step S110> In the acquisition step S110, evaluation information including oral cavity information is acquired. For example, the acquisition unit 11 may acquire the evaluation information by a different method for each type of information included in the evaluation information. That is, the acquisition unit 11 may acquire some of the information included in the evaluation information (e.g., oral cavity information) from the user terminal 2, and acquire other information (e.g., body information) from the input unit 108. In this case, it is not necessary to prepare all of the information included in the evaluation information every time fall information is generated, and it is possible to reduce the burden on the user.

[0061] <Generation step S120> In the generating step S120, the classification model is referred to and fall information corresponding to the evaluation information is generated. For example, the generating unit 12 may select a classification model suitable for the evaluation information from among a plurality of classification models based on basic information included in the evaluation information and generate the fall information by referring to the selected classification model. In this case, the fall information can be generated using a classification model suitable for the characteristics of an individual, which can lead to improved accuracy in estimating the possibility of falling.

[0062] For example, when a logistic regression model is used as the classification model, the generating unit 12 generates fall information using the results of assigning odds to each piece of information included in the evaluation information (for example, oral cavity information, subjective information, physical information, and cognitive information). The odds indicate the degree of contribution when estimating the possibility of falling. As the odds, for example, a value at which the objective variable becomes 1 when the values ​​of all explanatory variables are 0 may be used, or, for example, a change in the odds (odds ratio) at which the objective variable becomes 1 when the explanatory variable of interest increases by 1 with the other explanatory variables fixed may be used. This makes it possible to suppress variations in estimation accuracy caused by the type of information included in the evaluation information.

[0063] The support method of this embodiment is completed by performing each of the above steps. Note that at least any of the above steps may be repeatedly performed.

[0064] According to this embodiment, the acquisition unit 11 acquires evaluation information including oral cavity information related to the oral cavity function of the user. The generation unit 12 also refers to the classification model and generates fall information indicating the possibility of the user falling for the evaluation information. Therefore, the possibility of the user falling can be estimated based on the oral cavity information that is likely to affect falls. This makes it possible to improve the accuracy of estimating the possibility of the user falling.

[0065] According to the present embodiment, the evaluation information includes subjective information on the user's subjective health condition. Therefore, the possibility of the user falling can be estimated based on a combination of oral cavity information, which is likely to affect a fall, and the subjective information. This makes it possible to further improve the accuracy of estimating the possibility of the user falling.

[0066] Furthermore, according to this embodiment, the evaluation information includes physical information related to the user's physical functions and cognitive information related to the user's cognitive functions. Therefore, the possibility of the user falling can be estimated from multiple perspectives based on the four types of information that are likely to affect falls. This makes it possible to further improve the accuracy of estimating the possibility of the user falling.

[0067] Furthermore, according to this embodiment, the generation unit 12 generates fall information using the results of assigning odds indicating the degree of contribution of the oral cavity information, subjective information, physical information, and cognitive information included in the evaluation information to estimating the possibility of a fall. Therefore, the possibility of a user falling can be estimated based on the degree to which each piece of information contributes to the estimation of the possibility of a fall. This makes it possible to further improve the accuracy of estimating the possibility of a user falling.

[0068] Furthermore, according to the present embodiment, the generation unit 12 selects the proposed information based on the generated fall information. This can lead to a reduction in the possibility of the user falling. This makes it possible to improve the maintenance of the user's health.

[0069] According to the present embodiment, the acquisition step S110 acquires evaluation information including oral cavity information related to the oral cavity function of the user. The generation step S120 refers to the classification model and generates fall information indicating the possibility of the user falling for the evaluation information. Therefore, the possibility of the user falling can be estimated based on the oral cavity information that is likely to affect the fall. This makes it possible to improve the accuracy of estimating the possibility of the user falling. EXAMPLES

[0070] Next, examples related to the above-mentioned estimation system 100 and estimation method will be described. In the examples, first, as a first example, the relationship between the items included in the above-mentioned evaluation information and the fall outcome was evaluated. Then, as a second example, a classification model suitable for estimating a fall was derived and evaluated based on the items significantly related to the fall outcome.

[0071] (First embodiment) In the first example, responses to the 37 items shown in Figs. 5 to 8 were used to evaluate the relationship between each item and the outcome of the fall.

[0072] The lists shown in Figures 5 to 8 are the results of physical examinations and interviews conducted on the subjects. The total number of subjects (All) surveyed was 2,381, of which the responses of 1,234 subjects who had not experienced a fall within the past year (Falls = 0) and 449 subjects who had experienced a fall within the past year (Fall = 1) were used. The responses of 698 subjects who were determined to be invalid (NA) were not evaluated.

[0073] As for the answering method, for No.1-4 and 7-11, the corresponding numerical values ​​are used, and the list shows the average value (standard deviation in parentheses). For No.5 and 13-37, the answers to the questions ("0=No", "1=Yes") are used, and the list shows the number of answers (proportion in parentheses). For No.6, the answer to the gender is used, and the list shows the number of answers (proportion in parentheses). For No.12, the answers to the question about the subjective health condition ("6: Very bad", "5: Bad", "4: Not very good", "3: Average", "2: Fairly good", "1: Very good") are used, and the list shows the average value (standard deviation in parentheses). The "P-Value" value for each item indicates the P value for the tendency to fall for each item.

[0074] In the first embodiment, the data in the list above was used as the evaluation target, and an evaluation was performed using a generalized linear mixed model (GLMM) incorporating mixed effects in a multilevel logistic regression analysis. In this case, a classification model (logistic regression model) was derived using the answers to each of the above items as explanatory variables and the presence or absence of a fall experience as a response variable. The classification model was derived using the items No. A, No. B, and No. C described below in addition to the above items.

[0075] 9 and 10 show the results of evaluating the relationship between each item and the probability of falling using the above classification model. The responses of 1,504 subjects were used as the evaluation targets.

[0076] In the results shown in Figures 9 to 10, "Odds ratio" indicates the odds ratio for each item. "β" is the standard partial regression coefficient, and indicates the effect on the probability of falling. "p-value" indicates the p-value calculated by the Wald test, and the significance level was set to P<0.05.

[0077] In addition to the 37 items mentioned above, the evaluation targets include No. A, Living Area, N No.B was the duration of participation in a specific exercise (Participation; three levels: short-term, mid-term, long-term), and No.C was the interaction between age and duration of participation in a specific exercise (Interaction terms between age and participation; three levels: age: short-term, age: mid-term, age: long-term). The "intercept" is the probability of falling when the variables are centered (subtracting the average from the variable) at the average age of the target group and the value of each variable is set to "0". Centering is a method that does not affect the estimated value of "β" or the test results, and ensures that the "intercept" has an appropriate meaning. The "intercept" was calculated by simultaneously estimating the average effect of the entire sample and the variance component that varies probabilistically from it.

[0078] In the results of the first embodiment, the p-values ​​of items No. 10, 12, 19, 21, 24, 30, and B-3 (items in black in Fig. 9 and Fig. 10) were all less than 0.05. Therefore, the above items showed a tendency to be significantly related to the outcome of falls.

[0079] No. 10 (Example 1) shows the contents related to physical information. Example 1 showed a tendency that the probability of falling changes by -0.007 with an increase of one second in the time spent standing on one leg.

[0080] No. 12 (Example 2) shows the contents related to subjective information. Example 2 showed a tendency that the probability of falling changes by 0.147 with an increase of 1 point in the health condition.

[0081] No. 19 (Example 3) shows the contents related to physical information. In Example 3, when the subject answered "No" to the question "Can you usually stand up without an assistive device?", the probability of falling tended to change by 0.331 compared to when the subject answered "Yes".

[0082] No. 21 (Example 4) shows the contents related to physical information. In Example 4, when the subject answered "Yes" to the question "Have you experienced a fall in the past year?", the probability of falling tended to change by 1.149 compared to when the subject answered "No."

[0083] No. 24 (Example 5) shows the contents related to oral cavity information. In Example 5, when a person answered "Yes" to the question "Have you choked on tea or soup recently?", the probability of falling tended to change by 0.270 compared to when the person answered "No".

[0084] No. 30 (Example 6) shows the contents related to cognitive information. In Example 6, when the subject answered "Yes" to the question "Do you ever not know what date it is today?", the probability of falling tended to change by 0.210 compared to when the subject answered "No".

[0085] No. B-3 (Example 7) shows information on subjective information. Example 7 showed that when participants answered that they had been doing a particular exercise for a long time, the probability of falling tended to change by -0.261.

[0086] From the above evaluation, it was confirmed that Examples 1 to 7 were more significantly related to the outcome of falling than other items. Therefore, it was confirmed that it is possible to improve accuracy by estimating the possibility of a user falling using at least one of physical information, subjective information, oral cavity information, and cognitive information that are likely to affect a fall.

[0087] (Second Example) In the second example, two types of models were created, including those in Examples 1 to 7 described above. FIG. 11 shows the items used in creating the two types of models (Model 1, Model 2) and the evaluation results of each model. Model 1 was created using formula (1) shown in FIG. 11, and Model 2 was created using a formula in which unnecessary items were deleted from formula (1). As a result of evaluating the AUC (Area Under Curve) and sensitivity in each model, no significant differences were found, and therefore only the evaluation results of Model 1 are shown below.

[0088] FIG. 12 shows the results of calculating the AUC, cutoff value (Best cutoff), sensitivity estimate (Sensitivity), and specificity estimate (Specificity) for each of the 10 types of evaluation information (p1 to p10) using Model 1, and calculating the average value. As a result, the best cutoff value was 0.215. The sensitivity estimate (Sensitivity) was 61.0% to 82.6% (average 68.4%). The specificity estimate was 53.8% to 80.0% (average 64.2%). These results confirmed that Model 1 can be used to estimate the possibility of falling with sufficient accuracy.

[0089] It has been confirmed from the above-mentioned first and second examples that it is possible to improve the estimation accuracy by estimating the possibility of a user falling using evaluation information including oral cavity information.

[0090] Although an embodiment of the present invention has been described, this embodiment is presented as an example and is not intended to limit the scope of the invention. Such a novel embodiment can be implemented in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. This embodiment and its modifications are included in the scope and spirit of the invention, and are included in the invention and its equivalents described in the claims. [Explanation of symbols]

[0091] 1: Estimation device 2: User terminal 3: Communication network 4: Server 10: Housing 11: Acquisition part 12: Generation part 13: Output section 14: Storage section 100: Estimation system 101: CPU 102: ROM 103: RAM 104: Preservation Department 105: Interface 106: Interface 107: Interface 108: Input section 109: Display section 110: Internal bus S110: Acquisition step S120: Generation step

Claims

1. An estimation system for estimating a possibility that a user will fall, comprising: An acquisition unit that acquires evaluation information including oral cavity information regarding the user's oral cavity function; A storage unit in which a classification model derived using a plurality of pieces of reference evaluation information and a fall history linked to the reference evaluation information is stored; a generation unit that refers to the classification model and generates fall information indicating a possibility of the user falling in response to the evaluation information; To be prepared The estimation system comprises:

2. The evaluation information includes subjective information regarding the subjective health state of the user. The estimation system according to claim 1 .

3. The evaluation information is Physical information regarding the user's physical functions; and Cognitive information regarding the cognitive function of the user Contains The estimation system according to claim 2 .

4. the classification model represents a logistic regression model derived using the reference evaluation information as an explanatory variable and the fall history as a response variable, The generation unit generates the fall information by using a result of assigning odds indicating a degree of contribution of the oral cavity information, the subjective information, the physical information, and the cognitive information included in the evaluation information to estimating a possibility of a fall. The estimation system according to claim 3 .

5. The storage unit stores a plurality of pieces of proposal information regarding fall prevention measures linked to the possibility of falling, The generating unit selects the suggested information based on the generated fall information. The estimation system according to any one of claims 1 to 4,

6. 1. A method for estimating a likelihood of a user falling, comprising: An acquisition step of acquiring evaluation information including oral cavity information related to the user's oral cavity function; a generation step of referring to a classification model derived using a plurality of pieces of reference evaluation information and a fall history linked to the reference evaluation information, and generating fall information indicating a possibility of the user falling with respect to the evaluation information; To be prepared The estimation method is characterized by:

7. An estimation program for estimating a possibility that a user will fall, comprising: On the computer, An acquisition step of acquiring evaluation information including oral cavity information related to the user's oral cavity function; a generation step of referring to a classification model derived using a plurality of pieces of reference evaluation information and a fall history linked to the reference evaluation information, and generating fall information indicating a possibility of the user falling with respect to the evaluation information; To carry out An estimation program comprising:

Images