JP6474103B2 - Presentation device - Google Patents

Description

  The present invention is equivalent to what the administrator feels so that the health risk is not overlooked for the notification target person and is not taken too seriously when performing notification such as health guidance. The present invention relates to a presentation device that makes it possible to perform notification in a form that feels an appropriate risk.

  Conventionally, as shown in Patent Documents 1 and 2, etc., health guidance and alerting to the person have been performed for the purpose of reducing medical expenses.

JP 2014-225176 A JP 2014-225175 A

  However, in the conventional methods as described above, when trying to call attention, the two points (1) and (2) described below together with the background have not been addressed.

  (1) The Ministry of Health, Labor and Welfare has announced a policy of adding five mental illnesses to the so-called four major illnesses of cancer, stroke, acute myocardial infarction, and diabetes. It has become. However, mental health measures may feel that when the target group for implementing the measures is identified, it will affect the personnel assessments of companies regardless of whether or not they are healthy. Many have become psychological resistance.

  (2) In addition, when health guidance is given, people who have similar health conditions may be clustered for guidance. At this time, there is a problem that the cluster classification is not always optimal because it is not known at what granularity the cluster is most likely to cause behavioral change.

  In both cases (1) and (2), cognition that occurs when a healthy high-risk person group (cluster) to be instructed is presented becomes a problem.

  In the case of the former (1), when a person on the corporate side who provides health guidance is presented with the guidance target cluster, it is equivalent to a mental illness patient even for those who are not at risk of mental illness if they are properly managed. There is a possibility of evaluation.

  In the case of the latter (2), even if it is determined that you, your spouse, health insurance association, etc. belong to a specific group of high-risk individuals, you may overlook the risk without thinking it is too dangerous There is.

  In other words, when a specific label or cluster (healthy high-risk person in this example) is assigned to the target person, even if it is numerically dangerous or not dangerous, it does not agree with these judgments depending on cognition There is. In other words, the former (1) has an overestimation of risk, and the latter (2) has an underestimation of risk.

  The present invention has been made in view of the above-described problems of the prior art, and an object of the present invention is to provide a presentation device capable of notifying a person to be notified of an appropriate risk and making a notification about a health condition. .

  In order to achieve the above object, the present invention provides a presentation device, which is a collection of individual data for each subject and each age group, as well as health status data of a series of subjects given in the form of a bug of word. By applying latent topic analysis as a clustering, each clustering is performed at a predetermined series of granularities, and a clustering unit for obtaining a clustering result at each granularity, and an administrator determines which granularity of the clustering results of each granularity. Specifying whether to use the object for health guidance, and specifying which individual data of which cluster among the clustering results of the specified granularity are to be notified as health guidance targets. The manager for receiving a notification target receiving unit and the specified notification target person developing a specific disease A first risk value calculation unit that calculates a first risk value in a position based on the specified cluster, and the notification target person's position regarding the specified notification target person developing a specific disease A second risk value calculating unit that calculates a second risk value for each of the series of granularity clustering results, and a second risk value calculated for each of the series of granularity clustering results, A presentation determination unit that determines a clustering result at a granularity when it is determined to be closest to the risk value as a result for performing health guidance regarding the occurrence of a specific disease for the notification target person. Features.

  According to the present invention, the health guidance regarding the onset of a specific disease is performed based on the clustering result such that the notification subject feels the second risk value that is close to the first risk value as the objective value that the administrator feels. As a result, it can be determined automatically. Therefore, by performing the notification according to the automatically determined result, it is possible to make the person to be notified feel an appropriate risk and to promote an appropriate decision making regarding improvement of the health condition.

It is a functional block diagram of the presentation apparatus concerning one embodiment. It is a flowchart of operation | movement of the presentation apparatus which concerns on one Embodiment. It is a figure which shows the schematic example for demonstrating the step of FIG. It is a functional block diagram of a clustering part. It is a figure which shows the typical example of all the medical data input into a documentation part.

  FIG. 1 is a functional block diagram of a presentation device according to an embodiment. The presentation device 10 includes a clustering unit 1, a notification target reception unit 2, a first risk value calculation unit 3, a second risk value calculation unit 4, a database 5, and a presentation determination unit 6. In FIG. 1, for auxiliary explanation, lines (L1 to L7) are attached to some of the data lines between functional blocks (complex parts). The black circles on the data lines indicate that the flow of the same data diverges, and conversely, in the part that simply intersects without the black circles, the intersecting data lines are separate and independent. It shall be expressed.

  The outline of the processing contents of each unit in FIG. 1 is as follows. In the following, a formal description of each part in FIG. 1 will be given first, focusing only on the exchange of data handled by each part. After the formal description is finished, the substantial technical meaning will be described with reference to FIGS.

  The clustering unit 1 receives medical data (such as medical examination data and receipt data) as input, and outputs each result of clustering with a series of granularities (number of clusters) to each location. In other words, the information is output to the notification target receiving unit 2 as indicated by a line L1, is output so that it can be referred to an administrator who uses the presentation device 10 as indicated by a line L2, and the first as indicated by a line L3. Output to the risk value calculation unit 3 and output to the second risk value calculation unit 4 as indicated by a line L4. Details of clustering in the clustering unit 1 will be described later.

  As shown in line L5, the notification target accepting unit 2 employs any of the granularity clustering results obtained from the administrator by the clustering unit 1 to maintain the subject belonging to any cluster as healthy. An instruction on whether to set as a notification target regarding the state is received, the received information is output to the first risk value calculation unit 3 as indicated by a line L6, and the received information is output as a second line as indicated by a line L7. Output to risk value calculator 4.

  The first risk value calculation unit 3 calculates the first risk value based on the risk value of the notification target person received from the notification target reception unit 2 as evaluated from the viewpoint of the administrator, and then decides to present the first risk value Output to part 6.

  The second risk value calculation unit 4 evaluates the risk value of the notification target person received from the notification target reception unit 2 from the viewpoint of the notification target person itself (and the relatives of the notification target person and other related parties as described later). After calculating the second risk value, the second risk value is output to the presentation determination unit 6. Here, as will be described later, the second risk value is not calculated as only one value, but a series of values are calculated according to the clustering result obtained by the clustering unit 1, and each of them is a presentation determination unit. Output to 6.

  As described above, the first risk value calculation unit 3 calculates the first risk value from the viewpoint of the administrator, and the second risk value calculation unit 4 calculates the second risk value from the viewpoint of the notification target person. In the calculation, as will be described later, a method of evaluating each factor in the risk psychology framework is adopted, and each of the calculation units 3 and 4 is based on the manager's perspective and the notification subject's perspective under the risk psychology framework. After calculating each factor, the first risk value and the second risk value are calculated as the overall score of each factor.

  In order to perform the calculation, the calculation units 3 and 4 include a common module configuration as shown in FIG. That is, the first risk value calculation unit 3 includes a first fear factor calculation unit 31 that calculates a fear factor from the viewpoint of the administrator, and a first unknown factor that calculates an unknown factor from the viewpoint of the administrator. The calculation unit 32, the first reliability factor calculation unit 33 for calculating the reliability factor from the viewpoint of the administrator, and the fear factor from the viewpoint of the administrator calculated by each of the above units 31, 32, 33, unknown A first total calculation unit 34 that calculates a first risk value as a total score of sex factors and reliability factors.

  Similarly, the second risk value calculation unit 4 includes a second fear factor calculation unit 41 that calculates a fear factor from the viewpoint of the notification target, and a second risk factor calculation unit 41 that calculates an unknown factor from the viewpoint of the notification target. Two unknown factor calculation units 42, a second reliability factor calculation unit 43 that calculates a reliability factor from the perspective of the notification target person, and a notification target person's point of view calculated by each of the above units 41, 42, 43 A second total calculation unit 44 that calculates a second risk value as a total score of the fear factor, unknown factor, and reliability factor.

  The database 5 records in advance information necessary for the first risk value calculation unit 3 and the second risk value calculation unit 4 to calculate each factor for calculating each risk value. When the three, four, three modules 31, 32, 33, 41, 42, and 43 perform calculations, necessary information is used for reference. The details of the information recorded in the database 5 will be described together with the detailed description of the modules 31, 32, 33, 41, 42, 43.

  The presentation determination unit 6 prompts the notification target person to improve the behavior appropriately when notifying the health status of each notification target person received from the administrator by the notification target reception unit 2 (i.e. A mode of notification that can be motivated) is determined, and notification is actually performed in the determined mode. Here, as will be described in detail later, a series of second risk values calculated by the second risk value calculation unit 4 for each notification target person corresponds to each notification mode, and for each notification target person, The notification mode in which the second risk value closest to the first risk value received from the risk value calculation unit 3 is calculated is determined as the one that can promote appropriate behavior improvement and the like.

  In addition to the above-described embodiment, various other embodiments are possible for the determination, and an appropriate notification mode can be determined according to a policy set in advance by an administrator or the like. For example, in the above, it was determined that it corresponds to the second risk value that is “closest” to the first risk value. You may decide to be the “closest” one.

  FIG. 2 is a flowchart of the operation of the presentation apparatus 10 according to an embodiment. FIG. 3 is a diagram showing a schematic example for explaining the steps of FIG. Hereinafter, the details of each part in FIG. 1 will be described by explaining each step in FIG. 2 while referring to FIG. 3 as appropriate.

  First, as a premise for explaining each step in FIG. 2, the concept of risk psychology used as an evaluation method in the present invention in the form of the first risk value and the second risk value described above will be explained. .

  From the conclusion, the first risk value is a numerical value of the risk value felt by the manager, and corresponds to “risk evaluation by experts” in risk psychology. In addition, the second risk value calculated for a series of notification modes is a quantification of the risk value that the notification target person will feel when notification is actually made in each notification mode, Corresponds to “the sense of security felt by ordinary people” in risk psychology (risk assessment as the reverse of risk).

  In risk psychology, research on the discrepancy between the above "Risk assessment by experts" and "Reassurance felt by ordinary people" is made, and the basic elements of risk assessment by experts and the factors that support risk perception by ordinary people are different. Is heading. It is said that there are "frightening factor" and "unknown factor" as factors of risk recognition. In recent years, the “trust” of the information provider has also been regarded as important.

  Therefore, in the present invention, these factors are modeled for both the expert and the general public, and the subjective risk value is obtained as a function of these factors. Specifically, each modeled factor is calculated in the above-described modules 31, 32, 33, 41, 42, 43, and the first comprehensive calculation unit 34 of the first risk value calculation unit 3 uses the administrator (expert The first risk value is calculated as the subjective risk value, and the second risk value calculation unit 44 of the second risk value calculation unit 4 calculates the second risk value as the subjective risk value of the notification target person (general person). Furthermore, the presentation determination unit 6 determines a notification mode that does not cause a mismatch. Therefore, with a notification mode that makes the notification subject feel the second risk value that is as close as possible to the first risk value felt by the administrator, with appropriate strength (without giving excessive fear about the risk In addition, it is possible to encourage the person to be notified to change their behavior (without risking to neglect the risk).

  The relationship between the risk value and behavior change (that is, decision making) can be modeled as follows. According to the present invention, it is possible to promote behavioral change by evaluating a risk value according to the model.

  When a certain state of health (label) (cluster), for example, “preliminary diabetes group” is identified and presented to someone (person, family, etc.), emotions are generated. The decision is made as a result of the emotion. Decision making is, for example, lowering the personnel evaluation of a person who is determined to be a mental illness preparatory group. In other words, there are two states depending on whether or not decisions are made. For example, when the label “Diabetes Reserve” is given, the risk is underestimated, and when a significant measure is not taken, or when the label “psychiatric” is given, The risk may be overestimated and the personnel evaluation may be lowered. This is because there is a difference between the actual risk and the subjective risk value that the person feels.

  Wg is the lost profit / cost when the decision is made (when the personnel evaluation is lowered), wb is the lost profit / cost when the decision is not made (when the personnel evaluation is the same as before) If p is the probability of suspecting the subject's health (suspected that it is a mental illness and lowering the evaluation), the expected cost of each state is wg × p if the decision is made and the evaluation is lowered. Conversely, if the evaluation is maintained without making a decision, wb × (1-p). It is assumed that a decision is made based on the expected cost. Note that wg and wb become weights if the view is changed. If both weights are the same, it will be judged only by the magnitude of the suspicion probability. (In other words, if you want to avoid the occurrence (suspect) of bad emotions, you should just make this weight equal.) Here, consider the design of this probability p. The probability p is considered to be determined by the subjective risk value felt by the subject. When the subjective risk value is s_r, p = J (s_r) using an arbitrary function J.

  That is, since the subjective risk value has a one-to-one correspondence with the probability p of behavior change (decision making) by a function J (generally an increase function), the degree of promoting behavior change is quantified in the present invention. , Use the risk value. In addition, although said description made the case of the decision-making from the standpoint of an administrator (expert), it is the same also in the standpoint of a notification subject (general person). According to the present invention, the manager is encouraged to make a decision to improve the health status of the notified subject with the same strength as the decision to notify that the health status should be improved. be able to.

  Since the premise has been described above, each step of FIG. 2 will be described below.

  In step S1, as preparation for the operation of the presentation device 10, data collection necessary for the operation is performed, and then the process proceeds to step S2. The data collection may be performed manually by an administrator or the like, and a series of medical data to be clustered by the clustering unit 1 and data recorded in the database 5 are collected. Details of each collected data will be described later.

  In step S2, the clustering unit 1 performs clustering of medical data with a series of granularities preset by the administrator or the like, that is, a series of clusters, obtains a clustering result at each granularity, and then proceeds to step S3. Here, in each cluster of the clustering result, the onset rate of the specific disease that the administrator is trying to notify is also obtained, and the process proceeds to step S3.

  Although details of the clustering will be described later, individual samples to be clustered are obtained from a series of medical data input to the clustering unit 1, and data D reflecting the health state of each subject X at each age n (X, n), and a result of classifying these series of samples D (X, n) is obtained. The onset rate may be obtained from the actual value of the sample belonging to each cluster as the sample ratio (for example, the ratio of samples that develop diabetes after one year) that develops after specific years have passed in each cluster.

  In addition to determining the incidence rate as the actual value of each sample belonging to each cluster of the clustering result as described above, the incidence rate of each cluster is determined as a result of analyzing the clustering result by the administrator's hand. Also good. For example, after obtaining the transition probability (transition probability for each year) between each cluster of the clustering result, and determining the cluster (onset cluster) that is determined to have already developed from the clustering result, specify in each cluster The probability of transition to the onset cluster after the lapse of years may be defined as the onset rate.

  In FIG. 3, in [1], a schematic example of the clustering result by the clustering unit 1 in step S2 is shown, and here, a series of granularities (number of clusters), four, four, four, four, and five , Each clustering result is the result CR2 (cluster C21 and C22 two clusters), result CR3 (cluster C31, C32, C33 three clusters), result CR4 (cluster C41, C42, C43, 4 clusters of C44) and results CR5 (5 clusters of clusters C51, C52, C53, C54, C55). Note that [2], [3], [4-1], and [4-2] shown in FIG. 3 and described later are examples corresponding to the example of [1].

  In step S3, the administrator selects a clustering result having an appropriate granularity for performing notification regarding the health condition from the clustering results in the series of granularities obtained in step S2, according to the administrator's own judgment, After determining which cluster in the clustering result is to be notified of the health condition and determining which target person belonging to the determined cluster is to be notified, the process proceeds to step S4.

  When the clustering result to be used is determined by specifying the granularity and it is also determined which cluster is the notification target, basically all target persons belonging to the determined cluster are set as the notification target. However, only a part of the information may be targeted for notification based on the judgment of the administrator.

  When the administrator makes a series of selections and decisions based on his / her own judgment in Step S3, he / she can confirm the series of clustering results actually obtained in Step S2 and make the judgment in Step S3. That's fine. In FIG. 3, after confirming the series of clustering results CR2 to CR5 of [1] in [2], the result CR3 at the granularity = 3 is determined to be used for notification, and the result CR3 An example of determining a cluster C31 as a notification target from among the above is shown.

  In the confirmation, basically, in each cluster of the clustering result of each granularity, the incidence rate of the specific disease of the belonging sample is confirmed, and the actual data content of the belonging sample is confirmed. It is only necessary that the administrator himself / herself determines which is properly classified.

  In [2] of FIG. 3, as the confirmed example, the clusters having the highest incidence of specific diseases in the respective results CR2 to CR5 are C21 to C51, and the incidence is 30%, 40%, 60%, and 80%, respectively. The case of is shown in tabular form. As shown in gray in the table, the administrator adopts the result CR3 as the result for notification, and determines the cluster C31 having the highest incidence of 40% as the notification target. Such a judgment is made by the administrator himself / herself considering that the clustering is too rough in the result CR2 and the clustering is too precise in the results CR4 and CR5, etc. .

  In addition, the cluster C32 in addition to the cluster C31 in the result CR3 is determined as a notification target, and the cluster C41 in the result CR4 in addition to the cluster C31 in the result CR3 is also determined as a notification target. Although it is possible in step S3, even when the plurality is selected, the processing by the presentation device 10 when selecting one cluster in one clustering result (that is, the processing after step S4) is performed for the selected plurality. Notification can be made by carrying out the same in each case. Therefore, for the sake of simplification of description, the description will be continued assuming that the selection in step S3 is one cluster of one clustering result.

  In step S4, the first risk value calculation unit 3 calculates the first risk value for the notification subject belonging to the specific cluster in the clustering result of the specific granularity selected in step S3, and then the process proceeds to step S5.

  The calculated first risk value is a risk value from the viewpoint of the manager as described above. Details of the calculation will be described later. In the example of FIG. 3, an example in which the risk value calculated for the cluster C31 of the result CR3 is “60” is shown at the bottom of the table [2].

  In addition, in [2] of FIG. 3, when the risk value is calculated in the cluster C21, C41, C51 with the highest incidence for other results CR2, CR4, CR5 that were not selected for comparison, Is "50,70,80". In the judgment of the manager in step S3, the first risk value calculated by the first risk value calculation unit 3 for each cluster of each result may be referred to.

  As shown in FIG. 2, steps S5 to S8 form a loop of repetition processing, and the repetition processing is performed once for each notification target person determined in step S3. In step S5, an arbitrary one who has not completed the process is set as a process target from the notification target persons to be the target of the repetition process, and then the process proceeds to step S6.

  In step S6, the second risk value calculation unit 4 calculates each second risk value in the series of notification modes for the notification target person selected in the latest step S5, and then proceeds to step S6.

  Here, regarding the setting of a series of notification modes for calculating the second risk value, in the first embodiment, it can be set depending on which of the series of clustering results of the granularity performed in step S2 is notified. . In the second embodiment, in addition to the granularity setting in the first embodiment, the setting is further made depending on who is actually notified of the health condition of the notification target person (who is the target of the presentation). can do.

  Examples of the first embodiment and the second embodiment will be described after the description of the flow of FIG. Details of the calculation of the second risk value will be described later.

  In step S7, for the notification target selected in the latest step S5, which of the series of second risk values calculated in step S6 is closest to the first risk value calculated in step S4. (Which is the optimum result) is determined by the presentation determination unit 6, and the health condition is notified to the notification target person by the notification mode corresponding to the determined second risk value, and then the step Proceed to S8.

  In determining the optimum result, as described above, various policies may be applied, such as imposing a condition “within a second risk value greater than or equal to the first risk value” on the condition “closest”. .

  The notification may be made by notifying the health status of the subject in the form of text information or the like. If step S6 is performed in the first embodiment, the clustering result when the second risk value as the optimal result is given is notified of the incidence of the specific disease in the cluster to which the notification target person belongs. You just have to do it. In addition, if step S6 is performed in the second embodiment, in the clustering result when the second risk value as the optimal result is given, the incidence of specific diseases in the cluster to which the notification target person belongs May be presented to the person to be presented when the second risk value as the optimum result is given.

  Here, although a specific example will be described later, in the first embodiment, the person's health condition is notified directly to the notification target person by text information such as an e-mail. In the second embodiment, it is expected that direct notification to the notification target person will not have an effect of encouraging motivation for health improvement (as a result of confirming a series of second risk values in step S7) The relevant person (for example, spouse, etc.) of the notification target person is notified of the health condition of the notification target person, and at the same time, the relevant person is notified of the health condition of the notification target person. A request is made. The request may be made in the form of text information or the like.

  Therefore, in the second embodiment, the person to be notified is not directly notified of his / her health status by text information such as e-mail, but sent directly or verbally by another person. Receive notification of your health status by e-mail. As will be described later, the second embodiment has an effect of more urgently making a decision on health improvement to the notification target person in consideration of reliability factors in risk psychology.

  In step S8, it is determined whether or not the processing in steps S5 to S7 has been completed for all notification target persons determined in step S3. If completed, the flow in FIG. 2 ends and is completed. If not, the process returns to step S5, and an uncompleted notification subject is set as a process target, and then the processes after step S6 are continued.

  [3], [4-1], and [4-2] in FIG. 3 show examples in which steps S6 and S7 are performed in the first embodiment and the second embodiment, respectively.

  In [3] of FIG. 3, when the results of steps S1 to S4 are [1] and [2], a series of granularities 2, 3, 4, The second risk value calculated in step S6 for each of 5 is shown in a tabular form, and as shown in gray, in step S7, the notification target person A is notified of his / her health status by the result 4 of the granularity 4 An example of what is done is shown.

  That is, as shown in [3], the second risk value when notifying the notification target person A with the results CR2, CR3, CR4, CR5 is calculated as “41, 52, 61, 73”, respectively. ing. On the other hand, as shown in [2], the administrator feels a risk value of “60” by adopting the result CR3. It is ideal to give a notification in the form of “Mr. A has been determined to belong to cluster C31 with a diabetes incidence of 40%, so let's improve his lifestyle”. However, if notified in this way, the person A to be notified will only feel the risk value “52” as shown in [3], which is less than the risk value “60” felt by the administrator, The effect of prompting decision making is insufficient. Therefore, as a result of being able to feel the risk value “61” that is closest to (and more than 60) the risk value “60” felt by the manager, CR4 “A has a 60% incidence of diabetes. Notification is made in the form of “I will improve my lifestyle because it is determined to belong to cluster C41”.

  The above example is an example when the notification subject A belongs to the most dangerous (highest incidence) cluster C41 in the result CR4. In this case, the notification is as described above. “Mr. A has been determined to belong to cluster C41, which has a 60% incidence of diabetes, so let's improve his lifestyle.” If the notification target person A belongs to another cluster in the result CR4, the notification is made based on the cluster to which the notification target A belongs and the onset rate in the cluster. For example, if you belong to cluster C42 and the incidence is 50%, a notice that "Mr. A has been determined to belong to cluster C42 with an incidence of diabetes of 50%, so let's improve lifestyle" Will be made.

  Further, as will be described in detail later, the risk value as exemplified in [3] is calculated as a value based on the cluster to which the notification target person A belongs in each result CR2, CR3, CR4, CR5. Become.

  On the other hand, in [4-1] and [4-2] of FIG. 3, when the results of steps S1 to S4 are as shown in [1] and [2], a certain notification target person B is used according to the second embodiment. The second risk values calculated in step S6 for each of a series of granularities 2, 3, 4, 5 and a series of information presentation candidates are shown in tabular form. Here, a candidate for presenting a series of information is to first present the health status of the person to be notified B (Mr. B) to the candidate, and then to M It is a candidate that is supposed to be asked to explain the state. When the candidate is determined to be Mr. B, the health status is directly notified to Mr. B as in the case of the first embodiment.

  In the example of [4-1] and [4-2] in FIG. 3, when two persons, Mr. B and Mr. C who are relatives of Mr. B, are set as the candidate, the second series of candidates. An example where the risk value is calculated is shown. That is to say, in [4-1], the second risk value when notifying candidate B (notification subject person B himself) is notified of the health condition with a series of granularities 2, 3, 4 and 5 respectively. An example of “32, 39, 45, 50” is shown. Also, in [4-2], when notifying Mr. B via candidate C (person B's relative), the number of cases when notifying the health status with a series of granularities 2, 3, 4, 5 An example in which the two risk values are “41, 53, 62, 70” is shown.

  As shown in [2], the administrator feels a risk value of “60” by adopting the result CR3, so if you are, you should notify Mr. B directly and feel the same risk value. Ideally. However, as shown in [4-1], Mr. B cannot feel the risk value “60” no matter how the notification is made. For example, in the clustering result CR5 with the highest risk, even if you directly notify Mr. B, “B has been determined to belong to cluster CR51 with an incidence of 80%, so let ’s improve his health.” Only 50 "can be felt.

  Therefore, as shown in [4-2], a higher risk value can be felt by receiving notification from C, a relative. Specifically, as shown in gray, from the relative CR, the result CR4 says "Mr. B has been determined to belong to cluster CR41 with an incidence of 60%, so let's improve his health" By telling Mr. B himself, he can make Mr. B feel the risk value `` 62 '' closest to the risk value `` 60 '' felt by the management side, and promote decision making for health improvement it can.

  Note that the above examples of [3], [4-1], and [4-2] in FIG. 3 may be notified to the notification subject as is by the result CR3 selected by the administrator shown in [3]. Since the risk value felt by the person to be notified is insufficient for the risk value felt by the manager, it has been an example of a decision to make a notification that makes a higher original risk value feel. That is, as it is, it is an example of dealing with a case where the risk is underestimated by the person to be notified.

  In the same manner as this, and conversely, when the risk is overestimated, the present invention can deal with it in a unified manner. In other words, if the notification target person is notified as it is based on the result selected by the administrator, the notification target person may feel an excessive risk value higher than the risk value felt by the administrator. Is possible. Even in such a case, a similar process (step S7 in FIG. 2) of searching for one close to the first risk value from the series of second risk values calculated by the second risk value calculation unit 4 It is possible to perform notification in such a manner that the notification target person feels the original risk value (the risk value similar to that felt by the administrator).

  Hereinafter, details of calculation of the first risk value felt by the administrator and the second risk value felt by the notification target side, which enables automatic selection of an appropriate notification mode as in the example of FIG. 3 above, That is, details of the first risk value calculation unit 3 and the second risk value calculation unit 4 will be described.

  As described above, the calculation units 3 and 4 both calculate the subjective risk value in each position (the position of the manager and the position of the person to be notified) in the framework of risk psychology. In risk psychology, it is modeled as a subjective risk value s_r = f (F, U, T). Where F = frightness factor, U = unknown factor, T = confidence, and f is a predetermined function (however, it is an increasing function with respect to the frightening factor F and confidence T, and with respect to the unknown factor U is reduced in one embodiment. Function). As will be described later, in another embodiment, a decrease function of U is adopted as the predetermined function f when the specific condition is satisfied, and U is set as the predetermined function f when the specific condition is not satisfied. An increase function may be employed (that is, whether the function f becomes a decrease function of U or an addition function is switched depending on whether a specific condition is satisfied).

  That is, the first risk value calculation unit 3 calculates the subjective risk value s_r from the standpoint of the administrator and outputs it as the first risk value. For this reason, the first fear factor calculation unit 31 calculates the fear factor F from the manager's standpoint, and the first unknown factor calculation unit 32 calculates the unknown factor U from the manager standpoint. The first reliability factor calculation unit 33 calculates the reliability factor T in the standpoint of the above, and by applying the predetermined function f to each factor F, U, T in the standpoint of the administrator calculated above, The first comprehensive calculator 34 calculates the subjective risk value s_r = f (F, U, T) from the manager's standpoint and outputs it as the first risk value.

  Exactly the same, the second risk value calculation unit 4 calculates the subjective risk value s_r from the standpoint of the person to be notified and outputs it as the second risk value. Therefore, the second fear factor calculation unit 41 calculates the fear factor F in the position of the notification subject, the second unknown factor calculation unit 42 calculates the unknown factor U in the notification subject position, The second reliability factor calculation unit 43 calculates the reliability factor T from the standpoint of the notification target, and applies the predetermined function f to each factor F, U, T calculated from the standpoint of the notification target. As a result, the second comprehensive calculation unit 44 calculates the subjective risk value s_r = f (F, U, T) from the viewpoint of the notification target person and outputs it as the second risk value.

  As described with reference to FIG. 3 and the like, the second risk value is calculated as a value corresponding to each notification mode.

  The details of calculating each factor F, U, T in each position (administrator's position and notification target person's position) will be described below. (Note that the calculation subject of each factor is each module 31, 32, 33, 41, 42, 43 as described above, so for simplicity of explanation, mention of these calculation subjects is omitted and only the calculation method is used. When calculating from the notification subject's standpoint, each factor F, U, and T may be a value that also depends on the notification mode, and the details of the calculation will also be described. In each calculation, contents pre-registered in the database 5 may be used, and details thereof will be described.

(1) About the calculation of the fear factor F The fear factor F is explained by the cost when the event occurs (for example, the cost for yourself when the subject A becomes diabetic) and the actual statistical onset probability. I think that it will be done. This event is, for example, an event that “subject A becomes diabetic”, that is, an event that the administrator considers as a criterion for promoting health improvement.

  Specifically, it can be calculated as F = h_F (D, R, pr (cl)). The function h_F may be a predetermined function similar to f described above, and may be an increase function for D and pr (cl) and a decrease function for R. Here, each value D, R, pr (cl) input to the function h_F and its calculation method are as follows.

  (1-1) D is a quantification of the severity of a disease (a specific disease that the administrator is paying attention to promote health improvement). The severity D of the disease is calculated by calculating the medical cost when suffering from the specific disease (the higher the medical cost, the more serious the disease) m and the life expectancy when suffering from the specific disease y Thus, D = g_D (m, −y) may be calculated using an arbitrary increase function g_D. That is, the higher the medical cost m, the greater the severity D, and the smaller the life expectancy y (the greater the negative “-y”), the greater the severity D. Here, the medical cost m and the life expectancy y related to the specific disease use those recorded in the database 5 as pre-registration data.

  As described above, when calculating as severity D = g_D (m, -y), both m and y are fixed values given for a specific disease, but the medical cost m is a value from the standpoint of the administrator (for example, the burden of the company) (Medical expenses) and the value in the position of the person to be notified (medical expenses paid by the person at the medical institution, etc.) generally differ from each other. Therefore, severity D depends on the value of the administrator and the position of the person to be notified. Different values. The life expectancy y is a value that does not depend on the manager's position and the notification target person's position.

  Note that the severity g_D (m, -y, s) may be calculated in consideration of the degree of knowledge s regarding the specific disease. The degree of retention s is obtained by conducting a questionnaire for each individual in advance regarding the degree of understanding of a specific disease, etc. in advance for each individual, digitizing the results and recording them in the database 5 for each notification. What is necessary is just to be able to refer to the value in the subject. The fixing degree s from the standpoint of the administrator may be recorded in the database 5 as the maximum value (so-called perfect value) of the results and referred to as the value of the administrator.

  In addition, when the second embodiment described above is adopted for the second risk value, the above-described fixing degree s uses the value of the person to be presented instead of using the value of the person to be notified. Also good. The degree of establishment s for the specific disease of the person to be presented may be calculated by questionnaire results or the like and recorded in the database 5 as described above.

  (1-2) R is basically used when the second embodiment described above is applied to the second risk value, and the notification target person and the information presentee (for example, [4 in FIG. -2] is the social distance between principal B and his relative C).

  Here, by including in the definition of the distance R the case where the person to be notified matches the information presenter, both the first embodiment and the second embodiment can be dealt with. When calculating the first risk value, the distance R is the distance between the notification target person and the administrator, and when calculating the second risk value, the notification target person and a series of information recipients Each distance between the target person and the manager may be used.

  Specifically, the distance R can be calculated as R = g_R (r, d). Here, r is a distance reflecting a general relationship, and d is a distance reflecting the actual situation for each notification target person. The distance R may be calculated using only one of these.

  r is expressed as a numerical value in which a general relationship (for example, a general relationship such as “Principal and its spouse”, “Principal and their friend”) is fixed as a distance (ie, (symmetric) matrix fixed) Prepare in advance (as a table) and record it in database 5.

  Here, in general, the fixed table is prepared on the assumption that the distance r is smaller as the relationship is closer to the person, but it may be prepared in consideration of the following. In other words, it is a consideration of the cost difference between the information presenter and the person who did not take the health improvement action, so that the "person and spouse" (or family) A fixed table may be prepared so as to reduce the distance ")". (In other words, taking action to improve health itself has a cost (cost to take time to exercise, etc.) and it feels awkward, but when viewed from the family etc., the cost is small, and on the contrary A fixed table may be prepared that reflects the fact that the cost is higher if not taken, and [4-1] and [4-2] in FIG. Yes.)

  d is a database prepared in advance, calculated as the distance on the phone book on social networks, mobile phones, etc. (ie, the total distance of the paths to the target when creating a relationship network) Record it in 5. In addition, the existing technology can be used for the calculation of the distance on the relationship network.

  By using a predetermined increase function for g_R (r, d), the value of R is linked to the magnitude of the distance (R does not increase when the distance is small). it can. When calculating as a terrifying factor F = h_F (D, R, pr (cl, n)), the function h_F should adopt a decreasing function with respect to the variable R so that the value of F increases as the distance decreases. That's fine.

  (1-3) pr (cl) is an actual statistical incidence (probability) in the cluster cl including the person when the health condition of the notification subject is classified into clusters. This schematic example is as shown in the table of FIG. 3, and the actual value in each of the clustering results obtained by the clustering unit 1 (or the value obtained from the transition probability by analyzing the clustering result as described above). ), The probability pr (cl) can be obtained.

  In addition, as shown in the examples [2] and [3] in FIG. 3, in the present invention, there is a room for selection to notify the notification target person with another clustering result for the clustering result selected by the administrator. The fluctuation of the fear when notifying each result is automatically calculated based on the probability pr (cl).

  Note that this probability pr (cl) is less reliable when the number of people constituting the cluster is too small. Therefore, the number of clusters may be n, and the rewritten pr (cl, n) may be adopted. Specifically, pr (cl, n) may be prepared as a pre-setting that is smaller than the actual value pr (cl) by a certain percentage when the number n of clusters is small. For example, “pr (cl, n) = pr (cl) ד correction term depending on n ””, and the correction term is a correction term that makes the value of pr (cl) smaller when n is small. Design in advance.

(2) Calculation of unknown factor U In risk psychology, unknown factors differ depending on events. In the present invention, “a certain notification subject develops a specific disease” is adopted as an event. Therefore, the unknown factor U is determined by “objective literacy (degree of knowing) for the disease” and “objective literacy (degree of knowing) for the information subject to be notified”. To do. The unknown factor U is defined as the unknown factor U (the unknown level U) becomes smaller as each literacy (the level of knowledge) increases.

(2-1)
Examples of observed quantities (observed quantities used as input values for functions) for estimating literacy (L) for diseases are as follows.
・ Health literacy questionnaire score: q ・ Disease history (receipt information): rh (vector)
Therefore, L = g_L (q, rh). Health literacy can be manipulated to some extent by education.

  Here, both q and rh are acquired in advance for each notification target person and recorded in the database 5. When calculating the first risk value from the manager's standpoint, the highest value may be adopted as the score q, as in the case of acquiring the above-mentioned fixing degree s. Thereby, the second risk value felt by the notification subject can be reduced as the health literacy of the notification subject is lower than the maximum value. Therefore, the function g_L (q, rh) of the literacy L may be an increasing function with respect to the score q. Regarding the receipt information rh, the information on the subject of notification may be used without depending on the position of the administrator and the subject of notification, and the information on the healthy person who is not worried about the specific disease from the standpoint of the administrator. You may make it use. Then, the function g_L (q, rh) of the literacy L may be prepared on a rule basis with respect to rh so that the literacy L becomes smaller as the disease history rh becomes more serious.

(2-2)
Examples of observed quantities for estimating “literacy of the information presenter for the notification target” (referred to as O) are as follows. That is, as in the case of (1-2) above, if the observation amount (r, d) for the “relationship / distance between the notification subject and the information presenter” is obtained and registered in the database 5 and used, Good. By calculating O = g_O (r, d) using the decreasing function g_O, the relationship is closer, so the better the information presenter knows the notification target, the greater the literacy O I can do it. For example, the literacy O may be calculated as a decreasing function with respect to the distance R described above.

  As described above, the unknown factor U can be calculated as U = h_U (L, O) based on the observed quantities (2-1) and (2-2). h_U is a predetermined decreasing function. That is, as each literacy L, O is larger, the unknown factor U is smaller, and as a result, the subjective risk value s_r = f (F, U, T) is larger in an embodiment in which U is a decreasing function.

  Note that, as described above, in one embodiment of calculating the subjective risk value s_r = f (F, U, T) as a decreasing function with respect to the unknown factor U, the risk of a specific disease is neglected in the literacy of the notification target person This is a preferable embodiment when it is known in advance (from the standpoint of an administrator, etc.), and conversely, it is always preferable when the risk of a specific disease is overestimated in the literacy of the notification subject. Is not limited. Therefore, as an embodiment that makes it possible to provide appropriate notification in the presentation determination unit 6 regardless of whether the literacy of the notification target person is neglected or overestimated, the subjective risk value The function f of s_r = f (F, U, T) may be used separately for the unknown factor U as a decreasing function or an increasing function.

  Specifically, the case determination may be automatically performed by performing pre-calculation as in [Procedure 1] and [Procedure 2] below.

  [Procedure 1] The subjective risk value from the standpoint of the manager is calculated by s_r = f (F, U, T) by giving the unknown factor U a predetermined minimum value Umin as a temporary value. Calculate as the first risk value. Similarly, s_r = f (F, U, T) by giving a predetermined minimum value Umin as a tentative value, which is the same as the value in the administrator, to the unknown risk factor U as the subjective risk value in the notification subject position And calculated as a “provisional” second risk value.

  The minimum value Umin of the unknown factor may be a predetermined value given in advance. For example, each literacy L, O is given in advance as a maximum value Lmax, Omax (a common value for the administrator and the notification target person). The value Umin = h_U (Lmax, Omax) in the case of the temporary predetermined maximum value given as follows may be adopted. These values may be registered in the database 5.

  Here, in the calculation of the provisional first risk value and the provisional second risk value, a decreasing function of U may be used as the predetermined function f. Further, for F and T, which are arguments other than those given as U = Umin (virtually), actual values in the manager's position and the notification target person's position may be given. However, regarding the setting of the cluster for notification, only one type of provisional information regarding the notification target person can be obtained by adopting the same one cluster designated by the administrator received by the notification target reception unit 2 as it is. Find the second risk value of. (The provisional second risk value is not calculated for each series of clusters.)

  [Procedure 2] If the calculated first temporary risk value for the administrator is larger than the temporary second risk value for the notification target, it is possible to cope with the risk that the notification target neglects the risk. Therefore, the decreasing function of the unknown factor U is adopted as the predetermined function f. (When U decreasing function is adopted, U = h_U (L, O) is a decreasing function, so the function f is an increasing function of literacy L, O.) Conversely, the temporary first risk value is If it is less than or equal to the provisional second risk value, an increasing function of the unknown factor U is adopted as the predetermined function f so as to be able to cope with the risk that the notification subject overestimates the risk. (When the increasing function of U is adopted, since U = h_U (L, O) is a decreasing function, the function f is a decreasing function of literacy L, O.)

  As is clear from the above, the following is also possible as another embodiment. In other words, when it is known in advance (from the standpoint of the administrator, etc.) that the risk of a specific disease is overestimated in the literacy of the person being notified, the above [Procedure 1] and [Procedure 2] An increase function of the unknown factor U (that is, a decrease function of the literacy L, O) may always be adopted as the predetermined function f without dividing the case.

(3) About calculation of reliability factor T Quantification of reliability is generally difficult, but here it is defined from the relationship on the human relationship network, like distance R in (1-2). And T = g_T (r, d) may be calculated using information pre-registered in the database 5. g_T is a predetermined decreasing function. That is, as the distance R is smaller and the reliability factor T is larger, the subjective risk value s_r = f (F, U, T), which is an increasing function with respect to T, also becomes larger.

  When calculating the distance R on the social network with the existing technology, the calculation taking into consideration the communication performance etc. should be performed to contribute to the quantification of the reliability factor T. May be.

  Hereinafter, precautions related to effectively operating the presentation device 10 will be described.

  First, the “frightening factor” and the “unknown factor” in the above are the degree to which the person (the person to be notified) recognizes and feels, and the value of the “unknown factor” varies depending on literacy.

  In addition, if there is a discrepancy between the subjective risk value and the actual risk value (risk value from the manager's standpoint) for the notification target person, it is necessary to design a function so as to correct it to correct recognition. If the subjective risk value is too large, reduce the unknown factor or give a correct perception of cost. If the subjective risk value is too small, increase the unknown factor (provide correct recognition of what you think you know) or give correct recognition of cost.

  The points in the present invention are the event occurrence probability pr (cl) that changes the value of the above-mentioned fear factor after designing the function, and the relationship with the subject who affects all the factors (the person who communicates, the information presenter) ) To change the subjective risk value and the suspicion probability. A specific example is as shown in [3] of FIG. Furthermore, by providing disease literacy that affects unknown factors in advance, further improvement of the effect can be expected.

  In addition, there is a possibility that there is a possibility that no matter how much the cluster granularity is changed when notifying the person himself / herself, but the information is communicated via a reliable person once (the result is communicated to the family and the family In this method, the subjective risk value may change. Specific examples are as shown in [4-1] and [4-2] in FIG.

  Hereinafter, supplementary matters in the present invention will be described.

  (Supplement 1) Clustering in clustering unit 1

  FIG. 4 is a functional block diagram of the clustering unit 1 and includes a documenting unit 11 and a classification unit 12 as shown.

  The documenting unit 11 reads all medical data as input data for clustering by the clustering unit 1 (the classification unit 12), and documents of each target person X constituting the all data at each age n (age n) Generated medical data D (X, n) is output to the classification unit 12.

  Documenting to the medical data D (X, n) is converted into a well-known bug of words format, that is, a document vector format with the frequency (number of occurrences) of each predetermined word as an element. The data D (X, n) is a vector reflecting the health state of the subject X at the age of n. The documenting is performed as preprocessing for enabling clustering in the rear classifier 12. Specifically, it is as follows.

  First, all input medical data is an evaluation of the health status of a series of subjects at a series of times. Specifically, for example, the results of a health check conducted under a health association, etc. As a result of an inquiry, a receipt (medical remuneration statement), or a combination thereof can be used.

A predetermined plurality of m words i ₁ , i ₂ ,..., I _m that are described in the medical data or are known to be able to be described are prepared in advance, and the documenting unit by analyzing the text of the medical data in n age of the subject X at 11, the word _{i 1, i 2, ...,} i m vector D (X, n) representing the health condition as the frequency vector of be generated it can.

For example, if the word i _b corresponding to the name of a specific disease in medical examination data or the like exists in the medical data of the subject X at the age of n, the value of the element of i _b of the vector D (X, n) is set to “1”. If it does not exist, the value of the same element can be set to “0”. Similarly, the word i _b such as the prescribed drug name in the receipt data or the like can be set to “1” or “0” depending on whether or not the word exists. Further, if the word i _b appears one or more times in the medical data of the subject X at the age of n, the value of the element of the i _b may be the number of appearances (that is, the frequency).

  In addition, for items to be evaluated numerically, such as body weight and blood test results in health checkup data, a predetermined word corresponding to the item is prepared, and according to a predetermined rule (predetermined function, etc.) according to the evaluation numerical value The frequency of words can be calculated and used as the element value of the vector D (X, n). Regarding the conversion from the evaluation numerical value to the word frequency, Japanese Patent Application Laid-Open No. 2015-32013 (invention name: numerical data analysis device and program) and Japanese Patent Application No. 2013-163207 (numerical data analysis device and Program), Japanese Patent Application No. 2013-217817 (numerical data analysis apparatus and program) may be used.

  In addition to the case of numerical values (quantitative data) as described above, in the case of qualitative data (for example, the presence or absence of a smoking habit described in a questionnaire, etc.) To the value of the element of the vector D (X, n).

As described above, each of the words i ₁ , i ₂ ,..., I _m is a health condition evaluation item in the input medical data (not only a direct evaluation but also a drug name in the receipt data. Each of the predetermined rules (words i ₁ , i ₂ , ..., i _m) for the evaluation result of the subject X at the age of n In this case, the document unit 11 generates a document vector D (X, n).

  FIG. 5 shows a schematic example of all medical data input to the documenting unit 11. As shown in the example, it is assumed that all medical data as input has a defect. In other words, it is desirable that long-term data is originally available, but in practice, as in the example of FIG. 5, only data over a short period of several years can often be used.

  In the example of FIG. 5, for example, there is data for 40-year-old to 43-year-old for Mr. A. Therefore, D (A, 40), D (A, 41) ), D (A, 42), and D (D43) are output. For other subjects such as Mr. G and Mr. D, data corresponding to the age at which medical data exist will be output.

First, the classification unit 12 performs clustering on all the data D (X, n) of a series of ages n in the series of subjects X output from the documenting unit 11 described above. LDA (latent Dirichlet allocation method) based on the latent topic model can be used for the clustering. Note that LDA is disclosed, for example, in Non-Patent Document 1 below.
[Non-Patent Document 1] D. Blei, A. Ng, and M. Jordan. Latent Dirichlet allocation. Journal of Machine Learning Research, 3: 993-1022, January 2003.

  In particular, data D (X, n) and D (X, m) whose age is different from that of n years old and m years old (m ≠ n) even for the common target person X is clustered as separate data. . For example, the data D (A, 40), D (A, 41), D (A, 42), D (D43) in Mr. A's four different ages in the example of FIG. It becomes. Note that the data D (X, n), D (Y, m), where the target is different from X and Y (where age n and age m may be the same or different) are naturally classified as clustering targets. Become.

  In the present invention, the incidence rate in each cluster of the clustering result is used. For each data D (X, n) belonging to each cluster, calculation of a predetermined incidence rate such as “onset of diabetes after 1 year” is performed. A criterion is provided, and the onset rate may be acquired as the actual value in each cluster. That is, for the total number of samples in each cluster, the number of samples (number of onsets) that satisfies a predetermined criterion such as “I developed diabetes after one year” may be adopted as the onset rate. Further, as described above, an administrator or the like may calculate a transition probability with respect to the clustering result, thereby obtaining a value such as “probability of transition to a diabetes onset cluster after one year” or the like as an onset rate.

  (Supplement 2) The documentation unit 11 converts the input medical examination data and other medical data into data D (X, n) as a bug of word corresponding to the health state of each subject X in each age n As described above, the documenting unit 11 may be omitted when the input data is converted into the bug of word format in advance.

  (Supplement 3) A message for notifying the health improvement in the presentation determining unit 6 may be generated by preparing a predetermined text or the like as a template. For example, in the case of performing notification in the form of [2] in FIG. 3 with the result CR3, “Mr. A has been determined to belong to cluster C31 with a 40% incidence of diabetes, so let's improve his lifestyle”. For the “40%” part and “C31” part, a template in which a value (text) corresponding to the record is set is prepared, and a message corresponding to the record may be automatically generated.

  (Supplement 4) For each person to be notified, who should be a relative or other series of information recipients should be registered in advance in the list, etc. for each person to be notified and recorded in database 5. . Similarly, various kinds of information necessary for calculating the second risk value regarding the information presenter may be recorded in the database 5 in the same manner. Similarly to the above (Supplement 3), the text etc. for notifying the information recipient of the health condition of the person to be notified and requesting that the person to be notified to actually notify the person to be notified By creating a template, you can generate it automatically.

  (Supplement 5) The notification of the message as described in the above (Supplement 3) and (Supplement 4) may be notified by using the existing media such as e-mail in the presentation determination unit 6. Not only text information but also voice information automatically generated from the text may be notified, or notification may be notified by automatically generated moving image information. In addition, various notification media can be used.

  (Supplement 6) In the above (Supplement 1), the medical data to be used was explained from a formal point of view. However, as actual all data, for example, data for the most recent years up to the present time point when a specific health insurance was implemented Etc. can be adopted. The target of the administrator's notification of the health condition (that is, specified as the target in the notification target receiving unit 2) is the age n among the D (X, n) existing in all the data. What is necessary is to target the point in time or the most recent point at which data exists.

  (Supplement 7) Regarding the calculation of the series of second risk values in step S6 of FIG. 2, the first embodiment for calculating with the series of clustering results for the notification target person, and the presentation target person (notification target) in the first embodiment In the second embodiment, the calculation is performed by changing not only the person himself but also his relatives. As described in the example of [4-1] and [4-2] in FIG. 3, even if the action improvement effect is not expected by the notification in the first embodiment, the presentation target other than the person himself / herself The improvement of the behavior was expected by the second embodiment that allows the person to be notified through the person.

  That is, the second embodiment allows a case of notifying the person via a person other than the person to be notified, and by increasing the number of notification modes in the presentation determination unit 6, a series of second risk values in the first embodiment. As a result, it is possible to make notifications that can promote action improvement more effectively than in the first embodiment.

Exactly the same, a third embodiment will be described in which a series of second risk value calculation objects in step S6 in FIG. In the third embodiment, in addition to calculating only the health literacy L in the current status of the notification target person _{[current status]} , it is further assumed that the health literacy education was performed before notifying the notification target person. Literacy L improved by _{[After improvement]} is also calculated. Therefore, the variation of the notification mode is increased in that a prior literacy education can be added.

Here, the current literacy L _{[current status]} of each notification target person is the literacy L of the notification target person for calculating the above-described unknown factor U, and as described above, L = g_L (q, rh) It can be obtained from records in the database 5 or the like. Further, the improved literacy L _{[after improvement]} may be obtained assuming that the health literacy questionnaire score q becomes a predetermined value q _{[after improvement]} by prior education. That is, for each notification target, assuming that the current questionnaire score is q _{[current state]} and the score is q _{[after improvement]} in the prior education, the current state and the improved literacy may be calculated as follows.
L _{[Current status]} = g_L (q _{[Current status]} , rh)
L _{[After improvement]} = g_L (q _{[After improvement]} , rh)

In the above, the score q _{[after improvement]} improved by the prior education is calculated as a predetermined value or q _{[after improvement]} = E (q _{[current state]} ) with a predetermined function E modeling the literacy education effect. It is sufficient to register it in the database 5. In the above, the same value is used for the receipt information rh. However, as with the score q, the current score rh _[current] and the improved score rh _{[after improvement]} are used to determine the literacy L _{[current ] And} L _{[after improvement]} may be calculated.

Then, when the second risk value calculated with the _improved literacy L _{[after improvement]} is determined as the optimum value in step S7 in FIG. 2, the presentation determination unit 6 first applies literacy to the notification target person. Make sure you ’ve improved your literacy through education before giving notifications about your health. (Conversely, when the second risk value calculated in the _current literacy L _[current] is determined as the optimum value, the prior literacy education is omitted and the same notification as in the first embodiment is made.)

  Here, to provide literacy education and confirm that literacy has been improved by the education, for example, the presentation determination unit 6 sends a notification of attendance of a well-known e-learning to the notification target person by e-mail or the like. You can do it. Then, the e-learning server on the network provides the notification target person with educational materials related to the specific disease in the form of a web page, etc., and the notification target person passes the comprehension check test given as a choice problem on the web page. When it is determined that the literacy has been improved, the presentation determining unit 6 may further notify the health status of the notification target person. Other methods may be used.

As described above, according to the third embodiment, the same effects as those in the second embodiment can be obtained. For example, when [4-1] in FIG. 3 is the second risk value when the _current literacy L of the notification target person B _{[current state]} = 40 points, [4-2] Although it was an example of the 2nd risk value at the time of notifying via the presenter C, a 2nd risk value as shown to [4-2] can be made to feel also in 3rd embodiment. That is, in the third embodiment, e-learning or the like is given to the notification target person B, so that the literacy of the notification target person B is improved from, for example, the current literacy L _[current] = 40 points _{[after improvement ]} = 80 points, so even if the notification target person B is notified directly without the presenter C, the second risk when notifying with each clustering result as shown in [4-2] The value can be made to feel.

  (Supplement 8) The present invention can also be provided as a program that causes a computer to function as all of the units 1 to 6 of the presentation device 10 or any part thereof. The computer can adopt a known hardware configuration such as a CPU (Central Processing Unit), a memory, and various I / Fs, and the CPU executes instructions corresponding to the functions of the respective units of the presentation device 10. It becomes.

  DESCRIPTION OF SYMBOLS 10 ... Presentation apparatus, 1 ... Clustering part, 2 ... Notification object reception part, 3 ... 1st risk value calculation part, 4 ... 2nd risk value calculation part, 5 ... Database, 6 ... Presentation determination part

Claims (11)

By applying latent topic analysis as a collection of individual data for each subject and age, the health status data for a series of subjects given in the form of a bug of word is applied to a predetermined set of granularities. In each clustering, a clustering unit for obtaining a clustering result at each granularity,
The administrator specifies which granularity of the clustering results of each granularity is used for health guidance, and which individual data of which cluster among the clustering results of the specified granularity is healthy. A notification target reception unit that receives designation of whether to be a notification target person as an instruction target,
A first risk value calculation unit that calculates a first risk value in the manager's position regarding the specified notification subject developing a specific disease, based on the specified cluster;
A second risk value calculation unit that calculates a second risk value in the position of the notification subject regarding the specified notification subject developing a specific disease in each of the series of clustering results of the granularity;
Of the second risk values calculated for each of the series of granularity clustering results, the clustering result at the granularity when determined to be closest to the first risk value is the specific disease for the notification subject. A presentation determination unit configured to determine as a result for performing health guidance related to the onset of symptoms. In the first risk value calculation unit, the management for the specific disease given in advance as the first risk value in the manager's position regarding the specified notification subject developing the specific disease. Calculated based on the cost of
In the second risk value calculation unit, the second risk value in the position of the notification target person regarding the specified notification target person developing a specific disease, the second risk value for the specific disease given in advance The presentation apparatus according to claim 1, wherein the calculation is performed based on the cost in the position of the person to be notified. In the first risk value calculation unit, the notification target in the clustering result of the specified granularity is the first risk value in the manager's position regarding the specified notification target person developing a specific disease. Calculated based on the incidence of the specific disease in the cluster to which the person belongs,
In the second risk value calculation unit, a second risk value in the position of the notification target person regarding the specified notification target person developing a specific disease is calculated in each clustering result of the series of granularities. The presentation device according to claim 1 or 2, wherein calculation is performed based on an incidence rate of the specific disease in a cluster to which the notification target person belongs in a clustering result of each granularity. In the first risk value calculation unit, the management for the specific disease given in advance as the first risk value in the manager's position regarding the specified notification subject developing the specific disease. Calculated based on the literacy in the
In the second risk value calculation unit, the second risk value in the position of the notification target person regarding the specified notification target person developing a specific disease, the second risk value for the specific disease given in advance The presentation device according to claim 1, wherein the calculation is performed based on literacy from the standpoint of the person to be notified. In the first risk value calculation unit, the temporary first risk value is calculated by giving a temporary predetermined maximum value as the literacy value in the manager's standpoint, and in the second risk value calculation unit in advance , Calculating the provisional second risk value by giving the provisional predetermined maximum value as the literacy value in the position of the notification target person,
In the first risk value calculation unit and the second risk value calculation unit, as a function used when calculating the first risk value and the second risk value, respectively, the literacy in the position of the administrator and the notification target Using an increase function or a decrease function for literacy from the standpoint of the person who uses the increase function when the provisional first risk value is greater than the provisional second risk value, and the provisional first risk value is 5. The presentation device according to claim 4, wherein a decreasing function is used when the provisional second risk value is equal to or lower than the provisional second risk value. In the second risk value calculation unit, the second risk value in the position of the notification target person regarding the specified notification target person developing a specific disease, each of the clustering results of the series of granularity, The information presenter for performing information presentation regarding the occurrence of a specific disease for the notification target person,
In the presentation determination unit, out of the series of clustering results and the second risk values calculated for each of the information presenters, the clustering result at the granularity when determined to be closest to the first risk value and The information presenter is determined as a result for providing health guidance regarding the development of a specific disease to the notification subject, and the information presenter is developed to the specific notification subject. The presentation device according to claim 1, wherein the presentation device is determined as a person who presents information regarding what to do.   In the second risk value calculation unit, information on the specific disease given in advance as a second risk value in the position of the notification subject regarding the specified notification subject developing the specific disease The presentation device according to claim 6, wherein the calculation is performed based on the cost in the position of the person to be presented.   The cost in the information presenter's position for the specific disease given in advance is given based on the social distance reflecting the relationship between the information recipient and the notification subject. The presenting device according to claim 7, wherein the presenting device is provided. The information presentee includes the relative of the person to be notified, and the cost in the position of the information presenter as the relative is given in advance as a value higher than the cost in the position of the person to be notified. And
The presentation device according to claim 7, wherein the second risk value calculation unit calculates a second risk value in the position of the notification target person based on the cost given in advance.   In the second risk value calculation unit, the second risk value in the position of the notification target person regarding the specified notification target person developing a specific disease, and the information presenter who has been given in advance 10. The presentation apparatus according to claim 6, wherein calculation is performed based on reliability with the notification target person.   The reliability between the information recipient and the notification subject given in advance is based on the social distance that reflects the relationship between the information recipient and the notification subject. The presentation device according to claim 10, which is given as: