JP2016218966A - System and method for analysis - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a system and method for analysis which specifies a subject who implements a health support.SOLUTION: An analysis system includes a health transition prediction part 1308 which acquires health information of a candidate implementing an interposition, acquires time sequential health information from a starting time of the interposition of an individual implementing the interposition from a storage part as interposed health transition, predicts the time sequential health information of the candidate when the candidate starts the interposition as the interposed prediction transition based on the health information of the candidate and the interposed health transition and storing the predicted interposed predication transition in the storage part; a duration prediction part 1309 predicting the duration indicating the duration in which the candidate improves a health and storing the duration in the storage part; and a support target person selection part 1310 in which a processor selects the object person implementing the interposition from the candidates based on the duration.SELECTED DRAWING: Figure 1

Description

  The present invention relates to an analysis system and an analysis method.

  In organizations where multiple persons belong, such as local governments, companies, or insurers, the health status of members belonging to the organization is managed and health support (hereinafter referred to as health support) is provided to the members in order to maintain good health. It is important to implement. The reason for this is that maintaining good health is not limited to improving quality of life by preventing disease and increasing severity, but also has various effects such as reducing medical costs and maintaining labor force. is there.

  As a specific method of health support, for example, guidance for improving lifestyle habits can be mentioned. Even a person whose health condition is worsening can improve the future health condition and reduce the risk of illness if he / she can improve his / her lifestyle through guidance.

  On the other hand, when an organization actually implements such health support, there are constraints on resources such as the number of workers to be supported and expenses required for support. For this reason, organizations may not be able to provide health support for all those who need help. In such cases, organizations need to provide more effective and efficient health support to the appropriate audience while utilizing limited resources.

  As a system for implementing health support, the following methods have been shown so far.

  For example, in Patent Document 1, a means for calculating a guidance effect in time series for each group obtained by combining item conditions from medical examination information, guidance information, and receipt information, and an approximate straight line from a time series guidance effect. A health business support system having means for calculating as a feature quantity, means for selecting an item condition based on the feature quantity, and means for generating a health guidance model related to guidance priority using the selected item condition is shown. ing.

JP 2007-257565 A

  In order to provide effective and efficient health support with limited resources, it is important to provide appropriate health support for the target who is expected to be highly effective. The effect of support can be evaluated, for example, by comparing the health status before and after the implementation of support.

  On the other hand, although the effect of health support is great immediately after the support, it may be gradually weakened after a while after the support is provided. Therefore, in order to appropriately select the target person for health support, it is necessary to consider not only the health status before and after the implementation of support but also the long-term health status.

  Specifically, it is desirable to predict the long-term transition of the health condition after support, and to select subjects who can be expected to most sustain the effect of support.

  Therefore, an object of the present invention is to predict the long-term transition of the support effect and the duration of the support effect, and to appropriately select the target person for health support based on the predicted result.

  In order to solve the above-described problems, the present invention is an analysis system, which includes a processor and a storage unit, and the storage unit affects an index indicating an individual's health level and affects the individual's health. Storing health information including at least one of the indicators indicating factors that provide the health information and intervention information indicating the individual who performed the intervention supporting health, and the analysis system is a candidate for the processor to perform the intervention. The health information of the candidate, the time-series health information acquired from the start of the intervention of the individual who performed the intervention is acquired from the storage unit as a health transition with intervention, and the candidate's health information and the intervention Based on the health transition, the time-series health information of the candidate when the candidate starts the intervention is predicted as a predicted transition with intervention, and the predicted transition with intervention is stored in the storage unit. A health transition prediction unit, and the processor predicts a duration indicating a period in which the candidate's health is improved as an effect of the intervention using the predicted transition with intervention, and stores the duration in the storage unit And a processor for selecting the target person to perform the intervention from the candidates based on the sustain period.

  ADVANTAGE OF THE INVENTION According to this invention, the transition of the effect of health support can be estimated and the target person of health support can be selected appropriately based on a prediction result. Problems, configurations, and effects other than those described above will become apparent from the description of the following embodiments.

It is a block diagram which shows the structure of the analysis system of Example 1. FIG. 3 is a flowchart illustrating an outline of processing performed by a data analysis unit according to the first embodiment. It is explanatory drawing which shows the example of the diagnostic information of the health information storage part of Example 1, and lifestyle information. It is explanatory drawing which shows the example of the health assistance information of Example 1. FIG. It is a flowchart which shows the process by the health transition information generation part of Example 1. FIG. It is explanatory drawing which shows the example of the health transition information of Example 1. FIG. It is explanatory drawing which shows the example of the post-support transition information of Example 1. FIG. It is explanatory drawing which shows the example of the transition information without assistance of Example 1. FIG. It is a flowchart which shows the process by the health transition estimation part of Example 1. FIG. It is explanatory drawing which shows the health condition definition information of Example 1. FIG. It is explanatory drawing which shows the example of the support effect prediction transition of Example 1. FIG. It is explanatory drawing which shows the example of the prediction transition without assistance predicted of Example 1. It is a flowchart which shows the process by the duration estimation part of Example 1. FIG. It is explanatory drawing which shows the calculation method of the effect duration of Example 1. FIG. It is explanatory drawing which shows the example of the duration information of Example 1. FIG. 5 is a flowchart illustrating processing by a support target person selecting unit according to the first embodiment. FIG. 6 is an explanatory diagram illustrating an example of a user interface screen according to the first embodiment. 10 is a flowchart illustrating processing by a long-term support plan formulation unit according to a second embodiment. 10 is a flowchart illustrating processing performed by a health index narrowing unit according to the third embodiment. It is a flowchart which shows the process of the health transition estimation part of Example 4. FIG. 10 is an explanatory diagram schematically illustrating an example of a graphical model according to a fourth embodiment.

  Hereinafter, embodiments for carrying out the invention will be described with reference to the drawings.

  The provider provides health support to the individual by implementing an intervention plan that includes lifestyle guidance and instructions to the individual. However, there may be cases where the intervention plan cannot be implemented for all subjects who need health support due to restrictions on the number of leaders who implement the intervention plan and restrictions on expenses. In this case, it is desirable for the provider to implement the intervention plan preferentially for the target who can be more effective.

  On the other hand, the effectiveness of the intervention plan may gradually weaken after a while after the intervention plan is implemented. For this reason, in order to appropriately select subjects for health support, it is necessary to consider not only short-term effects based on comparison of health status before and after the implementation of the intervention plan, but also long-term effects. Specifically, it is desirable to predict the long-term transition of the effect of health support and select a target person who can expect the longest effect of health support as a target person who performs health support.

  In the first embodiment, an analysis system that predicts the duration of the effect of health support based on information on health and information on past health support, and selects a target person for health support based on the predicted duration. An example will be described. In addition, although health support in a present Example is performed by implementing an intervention plan, it describes as what implements health support only below.

  FIG. 1 is a block diagram illustrating the configuration of the analysis system according to the first embodiment.

  The analysis system according to the first embodiment includes a data analysis unit 1301 and a database 1350. The data analysis unit 1301 is a computer having an input unit 1302, an output unit 1303, an arithmetic device 1304, a memory 1305, and a storage medium 1306.

  The input unit 1302 is a human interface such as a mouse and a keyboard, and receives input to the data analysis unit 1301.

  The output unit 1303 is a display or a printer that outputs a calculation result by the analysis system.

  A storage medium 1306 is a storage device that stores a program for realizing data analysis processing by the data analysis unit 1301, an execution result of the data analysis processing, and the like. The storage medium 1306 is, for example, a nonvolatile storage medium (such as a magnetic disk drive or a nonvolatile memory).

  The memory 1305 includes a ROM that is a nonvolatile storage element and a RAM that is a volatile storage element. The ROM stores an immutable program (for example, BIOS). The RAM is a high-speed and volatile storage element such as a DRAM (Dynamic Random Access Memory), and temporarily stores a program stored in the storage medium 1306 and data used when the program is executed.

  The arithmetic device 1304 realizes the function of the data analysis unit 1301 by loading a program included in the storage medium 1306 into the memory 1305 and executing the program. The arithmetic device 1304 is, for example, a CPU or a GPU. The processing device 1304 executes processing and calculation described below.

  The storage medium 1306 executes the functions of a health transition information generation unit 1307, a health transition prediction unit 1308, a duration prediction unit 1309, a support target person selection unit 1310, a health index narrowing unit 1311, and a long-term support plan formulation unit 1312. Have a program to do.

  The health transition information generation unit 1307 generates health transition information based on the time series transition of health information and the implementation history of support. Based on the health transition information, the health transition prediction unit 1308 predicts the transition of health information with and without health support.

  The duration prediction unit 1309 predicts the duration of the health support effect. The support target person selecting unit 1310 selects a target person who performs health support.

  The health index narrowing-down unit 1311 does not execute the process described in the first embodiment, but executes the process shown in the third embodiment. The long-term support plan formulation unit 1312 does not execute the process described in the first embodiment, but executes the process shown in the second embodiment. For this reason, the storage medium 1306 according to the first embodiment may not include the health index narrowing unit 1311 and the long-term support plan formulation unit 1312.

  In this embodiment, the function of the data analysis unit 1301 is implemented by a program, but the data analysis unit 1301 has a physical integrated device that executes processing of each program, and the function is implemented by these integrated devices. May be. 1 may be executed by a plurality of programs or a plurality of integrated devices, and the plurality of functions may be executed by a single program or a single integrated device.

  The analysis system according to the first embodiment is a computer system configured on a single computer or a plurality of computers configured logically or physically, and operates on separate threads on the same computer. Alternatively, it may operate on a virtual machine constructed on a plurality of physical computer resources.

  The program stored in the storage medium 1306 is provided to the data analysis unit 1301 via a removable medium (CD-ROM, flash memory, etc.) or a network. For this reason, the data analysis unit 1301 of this embodiment may include an interface for reading a removable medium.

  The database 1350 includes a health information storage unit 1351, a health transition information storage unit 1352, a prediction information storage unit 1353, a health support information storage unit 1354, a duration storage unit 1355, a health condition definition storage unit 1356, a calculation index storage unit 1357, and The health transition prediction model storage unit 1358 is included.

  The calculation index storage unit 1357 is used in the third embodiment. The health transition prediction model storage unit 1358 is used in the fourth embodiment. Hereinafter, information on health (health information) handled in Example 1 and information on past health support (health support information 501 described later) will be described.

  The health information storage unit 1351 stores the health information input to the input unit 1302. The health support information storage unit 1354 stores information related to past health support input to the input unit 1302.

  The program of the data analysis unit 1301 acquires the contents stored in the health information storage unit 1351 and the health support information storage unit 1354 from another system or the like via a removable medium or a network, and moves to and stores the database 1350. May be.

  FIG. 2 is a flowchart illustrating an outline of processing performed by the data analysis unit 1301 according to the first embodiment.

  First, the health transition information generation unit 1307 generates health transition information based on the data stored in the health information storage unit 1351 and the health support information storage unit 1354 (2201). The health transition information may be stored in advance in the health transition information storage unit 1352, and the data analysis unit 1301 may not execute step 2201.

  After step 2201, the health transition prediction unit 1308 displays the transition of health information when the candidate subject to perform health support temporarily performs health support, the health transition information in the health transition information storage unit 1352, Prediction is performed based on the subject's health information (2202).

  After step 2202, the duration prediction unit 1309 predicts the duration of the health support effect based on the predicted health information transition of the candidate subject (2203). After step 2203, the support target person selecting unit 1310 selects a target person to perform health support from candidate target persons based on the predicted duration (2204).

  FIG. 3 is an explanatory diagram illustrating an example of the diagnosis information 301 and the lifestyle information 401 of the health information storage unit 1351 according to the first embodiment.

  The health information storage unit 1351 stores health information. The health information includes information related to personal health such as an index indicating the degree of personal health and life information indicating factors affecting the personal health. The health information is information acquired in advance from an individual through a medical examination, a test, an inquiry, and the like.

  The health information of the present embodiment includes diagnostic information 301, lifestyle information 401, and the like. The health information is stored in the health information storage unit 1351 via the input unit 1302.

  The health information storage unit 1351 in the present embodiment, when an individual provides health support, includes the timing when the individual starts health support (including immediately before or after starting health support) and after the individual provides health support. The health information acquired from the individual is included in at least three timings. The health information storage unit 1351 includes health information acquired from the individual at least at three different timings when the individual has not started health support.

  This is because the data analysis unit 1301 predicts the transition of health information based on the health information acquired at a plurality of timings in the processing described later.

  The diagnosis information 301 includes information indicating an individual's health level, and is information that holds a correspondence relationship between an individual and a result of a health check or the like. The diagnosis information 301 includes an individual ID 302, an acquisition year 303, an age 304, and various test value items (BMI 305 and fasting blood glucose 306) acquired at a medical institution or the like. In this embodiment, the index indicating the degree of personal health is, in particular, the items of age 304 and test value.

  The personal ID 302 is an identifier for uniquely identifying an individual. The acquisition year 303 indicates the year in which the results of the medical examination and examination conducted by the individual indicated by the individual ID 302 are acquired.

  The age 304 is the age when the result of the health check or the like of the individual indicated by the personal ID 302 is acquired. The BMI 305 and the fasting blood glucose 306 indicate the BMI and fasting blood glucose obtained in a health examination or the like performed by the individual indicated by the individual ID 302.

  The diagnostic information 301 may include any result other than that shown in FIG. For example, the diagnosis information 301 may include values such as a test value of a test that is generally widely performed in a health checkup and an index related to exercise ability acquired by an exercise test.

  The lifestyle information 401 includes lifestyle information that affects an individual's health, and maintains a correspondence relationship between the individual and information related to the lifestyle of each individual. Information on lifestyle habits is a result obtained by an individual answering a question, for example, a result of an inquiry.

  The lifestyle information 401 includes a personal ID 402, a reply year 407, a reply date 403, and reply contents (for example, cigarette 404, drinking 405, and walking 406). It should be noted that the lifestyle information 401 may include a lifestyle, a past history, a treatment history, a constitution such as allergy, subjective symptoms, etc. in addition to the response content shown in FIG. Further, the lifestyle information 401 may include information on a living environment that affects personal health, such as whether the land in which it lives is cold or warm.

  The personal ID 402 is an identifier for identifying an individual who has answered a question regarding lifestyle, and identifies the individual by the same identifier as the personal ID 302 of the diagnostic information 301. The reply year 407 indicates the year when the individual indicated by the personal ID 402 answers a question regarding lifestyle. The answer date 403 is the date when the question about the lifestyle habit is answered.

  The cigarette 404, drinking 405, and walking 406 are the answer contents obtained by the answer. The cigarette 404 indicates the average number of cigarettes smoked per day if the individual has a smoking habit, and indicates “none” if the individual does not smoke. The drinking 405 indicates an average daily drinking amount (unit = ml) when the individual has a drinking habit, and indicates “none” when the individual does not have a drinking habit. Walking 406 shows the average daily walking time (unit = minute) by an individual.

  In some cases, detailed information such as the number of steps, the amount of alcohol consumed, and the number of smokers cannot be obtained from individual responses. In such a case, the lifestyle habit information 401 may include a result of answering a corresponding one of the frequencies classified in advance in the questionnaire, instead of a specific amount of drinking.

  For example, the response content may indicate only the presence or absence of smoking or drinking. In addition, the response content may be a result of answering a plurality of classifications according to the frequency of drinking (for example, (A) no drinking, (B) 1-2 times per week, (C) 3 times per week). . In this case, the response content of the lifestyle information 401 indicates an identifier such as a number having no quantitative meaning.

  In addition, when an individual does not answer a specific item in a question, the data of the answer content may be lost. For example, in the lifestyle habit information 401 shown in FIG. 3, the lifestyle habit information of the individual whose personal ID 402 is “0003” lacks data for the walking 406 in the interview results answered in FY2004. In this case, the response content item may store a predetermined identifier indicating no response, such as a null value.

  The reply content of the lifestyle habit information 401 is an index indicating the degree of the influence of the matter affecting the individual's health and the presence or absence of the influence.

  Next, the health support information 501 stored in the health support information storage unit 1354 will be described.

  FIG. 4 is an explanatory diagram illustrating an example of the health support information 501 according to the first embodiment.

  Health support information (intervention information) 501 indicates an individual who has performed health support. The health support information 501 includes a personal ID 502, health support 503, year 507, start date 504, support status 506, and the like.

  The personal ID 502 is an identifier for identifying an individual who has provided health support in the past, and identifies the individual by the same identifier as the personal ID 302 of the diagnostic information 301. The health support 503 indicates the name of the health support that has been implemented. Further, the health support 503 may indicate a symbol or number such as an ID that can uniquely identify the health support, instead of the name of the health support.

  Year 507 indicates the year in which health support started. The start date 504 indicates the date on which health support was started. In this embodiment, health support is performed every year. When one health support is provided for one individual over a plurality of years, the health support information 501 includes entries for the same personal ID 502 and the same health support 503 for a plurality of years.

  The support status 506 indicates the implementation status of health support. The support status 506 indicates “finished” when the support is normally completed, and indicates “suspended” when the support is interrupted due to circumstances. When the support is continuing, the support status 506 may indicate “continuation”.

  Next, processing of the health transition information generation unit 1307 will be described. Based on health information such as diagnosis information 301 and lifestyle habit information 401 and health support information 501 that is information on past health support, the health transition information generation unit 1307 performs time-series transition of health information and implements support. Is generated in time series (health transition information 701 to be described later).

  FIG. 5 is a flowchart illustrating processing performed by the health transition information generation unit 1307 according to the first embodiment.

  The health transition information generation unit 1307 acquires information from the diagnostic information 301 and the lifestyle information 401 stored in the health information storage unit 1351 in accordance with the instruction input to the input unit 1302, and stores the information in the memory 1305 or the storage medium 1306. (101). The instruction input here may be, for example, a personal ID for identifying an individual, or a year for identifying entries of diagnostic information 301 and lifestyle information 401 (acquisition year 303 and answer year 407). Range).

  After step 101, the health transition information generation unit 1307 acquires the health support information 501 stored in the health support information storage unit 1354 according to the instruction input to the input unit 1302, and stores it in the memory 1305 or the storage medium 1306. (102). The instruction input here is the same as the instruction input in step 101.

  After step 102, the health transition information generation unit 1307 shows the health-related time-series transition of the health information for each individual based on the health information acquired in step 101 and the health support information 501 acquired in step 102. Transition information 701 is generated (103).

  FIG. 6 is an explanatory diagram illustrating an example of the health transition information 701 according to the first embodiment.

  The health transition information 701 illustrated in FIG. 6 is information stored in the health transition information storage unit 1352. The health transition information 701 is a result of summing up and integrating the health information and health support information 501 for each individual and for each predetermined period, and shaping the information into a table format.

  Here, the predetermined period is shorter than the range of the period input in step 101 and longer than the units of the acquisition year 303, the answer year 407, and the year 507. In the following description, the predetermined period is assumed to be one year, but it may be a period such as two years or three years.

  The health transition information 701 includes a personal ID 702, a year 703, a diagnosis result (BMI 704, fasting blood glucose 705), a health support 706, and a number of years 707 after the health support. The personal ID 702 is an identifier for identifying an individual, and identifies the individual with the same identifier as the personal ID 302 of the diagnostic information 301.

  The year 703 indicates a predetermined period, and is a value of the acquisition year 303, the answer year 407, and the year 507 used for sorting the diagnosis information 301, the lifestyle information 401, and the health support information 501.

  In step 103, the health transition information generation unit 1307 has the personal ID 702 and year 703 the same diagnosis information 301 entry as the personal ID 302 and acquisition year 303, and the personal ID 702 and year 703 the same lifestyle information as the personal ID 402 and answer year 407 One entry of the health transition information 701 is generated by merging the entry 401 and the entry of the health support information 501 whose personal ID 702 and year 703 are the same as the personal ID 502 and year 507.

  When the predetermined period is 2 years, the year 703 may indicate “from 2000 to 2001” or the like.

  The diagnosis result (BMI 704, fasting blood glucose 705) stores the value of BMI 305 and the value of fasting blood glucose 306 acquired in the year 703 from the individual indicated by the personal ID 302. The diagnosis result in the health transition information 701 stores a value obtained by tabulating the diagnosis information 301 for each individual ID 302 and for each predetermined period (year).

  In this embodiment, since the year 703 and the acquisition year 303 are the unit (year) of the same period, the health transition information generation unit 1307 sets the same values as the values of the BMI 305 and the fasting blood glucose 306 in step 103 to the BMI 704 and Store in fasting blood glucose 705.

  However, when the predetermined period in FIG. 6 is longer than the unit of the acquisition year 303 or the answer year 407, that is, when the period indicated by the year 703 is longer than the unit of the acquisition year 303 or the answer year 407, the health transition information generation unit 1307, at step 103, at least one of the statistically calculated values such as the average value, median value, mode value, and deviation of the diagnostic results (BMI 305 value and fasting blood glucose level 306) included in the predetermined period, The value of BMI 704 and fasting blood glucose 705 may be stored.

  Further, the confidence interval width may be a confidence interval width other than the 95% confidence interval width. For example, when the predetermined period is two years, the health transition information generation unit 1307 calculates the average value of the BMI 305 and the average value of the fasting blood glucose 306 from the entries of the diagnostic information 301 for two years, and the BMI 704 and the fasting rate. It may be stored in the blood glucose 705.

  The health transition information generation unit 1307 may further merge information such as lifestyle information 401 with the health transition information 701.

  The health support 706 indicates information related to health support performed by the individual indicated by the individual ID 302 in the year 703. In step 103, the health transition information generation unit 1307 stores information that summarizes the values of the health support 503 of the health support information 501 for each personal ID 302 and for each predetermined period.

  For example, in step 103, the health transition information generation unit 1307 is an entry including the same personal ID 502 as the personal ID 702, and the health support information 501 entry in which the year 507 is included in a predetermined period is stored in the health support information storage unit 1354. If specified, the health support name included in the health support 503 of the specified entry is stored in the health support 706. When the entry cannot be specified, the health transition information generation unit 1307 stores a value indicating “none” in the health support 706.

  In step 103, the health transition information generation unit 1307 may store, in the health support 706, ID information such as a number uniquely assigned to each health support in addition to the name of the health support. In addition, the health transition information generation unit 1307 may store, in the health support 706, information hierarchized according to the support status 506 information of the health support information 501.

  For example, the health transition information generation unit 1307 stores “intervention A (end)” in the health support 706 when the health support 503 indicates “intervention A” and the support status 506 indicates “end”. When the support 503 indicates “intervention A” and the support status 506 indicates “interruption”, “intervention A (interruption)” may be stored in the health support 706. Thereby, the health support 706 can indicate the implementation status of the support.

  The number of years since health support 707 indicates the number of years that have passed from the year in which health support was implemented to the year indicated by the year 703. In step 103, the health transition information generation unit 1307 calculates a value of the number of years after health support 707 based on the health support 706 and the year 703 of the entry with the same personal ID 702.

  For example, the health transition information 701 illustrated in FIG. 6 includes entries 708 and 709 whose personal ID 702 is “0003”. The entry 708 indicates that the individual with the personal ID “0003” performed health support for Intervention B in 2000. An entry 709 indicates that the individual with the personal ID “0003” has not started health support in FY2001.

  Therefore, the health transition information generation unit 1307 stores “0” in the post-health support elapsed years 707 of the entry 708 in step 103. Also, “1”, which is the difference between the year 703 of the entry 708 and the year 703 of the entry 709, is stored in the number of years since health support 707 of the entry 709.

  If there is no entry indicating that the same personal ID has performed health support in the entry of the health transition information 701 before the period indicated by the year 703, or the health support information 501 acquired in step 102 provides health support. If the entry indicating that the operation has been performed is not included, the health transition information generation unit 1307 stores, for example, “-” or a null value in the post-health support elapsed year 707 in step 103.

  The health transition information generation unit 1307 stores the health transition information generated by the above method in the health transition information storage unit 1352.

  After step 103, the health transition information generation unit 1307 shows information indicating the health information at the timing when the health support is started and the time series transition of the health information after starting the health support (hereinafter referred to as the post-support transition information 1401). Is generated based on the health transition information 701 (104), and further, information on the time series transition of the health information when the health support is not performed (the transition information without support 1501) is based on the health transition information 701. Generate.

  In step 104, the health transition information generation unit 1307 extracts information related to the individual who performed each health support from the health transition information 701 generated in step 103, and generates post-support transition information 1401.

  Specifically, the health transition information generation unit 1307 extracts an entry of health transition information 701 in which health support (hereinafter, intervention A) to be extracted is included in the health support 706. The extracted entry is an entry including health information of the year when the intervention A is started, and indicates health information after the start of health support.

  Next, the health transition information generation unit 1307 selects two or more entries having a post-health support age 707 of one or more of the health transition information 701 including the same personal ID 702 as the individual ID 702 included in the extracted entry. Extract. The entry extracted here is an entry including a transition of health information from the next year after the intervention A is performed, and indicates a time series transition of the health information after the health support is performed.

  Hereinafter, a method of generating post-supporting transition information 1401 from these extracted entries will be described with reference to FIG.

  FIG. 7 is an explanatory diagram illustrating an example of post-support transition information 1401 according to the first embodiment.

  In the following, a process for generating post-support transition information 1401 related to “intervention A” as in the above example will be described.

  Post-support transition information (health transition with intervention) 1401 indicates health information of individuals who have performed health support in chronological order. The post-support transition information 1401 includes health support 1402, years after health support 1403, personal ID 1404, year 1405, BMI 1406, fasting blood glucose 1407, and information retention years 1408.

  Health support 1402, years since health support 1403, personal ID 1404, year 1405, BMI 1406, and fasting blood glucose 1407 are health support 706 in health transition information 701 (FIG. 6), years after health support 707, personal ID 702, Information extracted from the year 703, BMI 704, and fasting blood glucose 705.

  The health transition information generation unit 1307 extracts entries including the health information of the year in which the intervention A was extracted, and the transitions of health information in the following year after the intervention A extracted from the health transition information 701 by the method described above. Are sorted in the order of health support 706, years after health support 707, and personal ID 702, and post-support transition information 1401 is generated.

  The information retention year 1408 indicates the number of entries in the post-support transition information 1401 having the same personal ID 1404. For example, the post-support transition information 1401 shown in FIG. 7 includes an entry indicating that the health support of the intervention A was implemented in 2000 for the individual whose personal ID 1404 is “0001”, and the 2000, 2001, And it includes entries for the year 2002 and a total of 3 years. Therefore, the health transition information generation unit 1307 stores “3” in all information holding years 1408 whose personal ID 1404 is “0001” in step 104.

  Further, the post-support transition information 1401 shown in FIG. 7 includes an entry indicating that the health support of intervention A was implemented in 2000 for the individual whose personal ID 1404 is “0005”, and 2000 and 2001 and 2 years in total. Includes minutes entry. Therefore, the health transition information generation unit 1307 stores “2” in all the information holding years 1408 whose personal ID 1404 is “0005” in Step 104.

  The health transition information generation unit 1307 can calculate the value of the information retention year 1408 from the maximum value for each individual ID 702 of the post-health support age 707 included in the health transition information 701.

  The health transition information generation unit 1307 stores the generated post-support transition information 1401 in the health transition information storage unit 1352.

  In step 105, the health transition information generation unit 1307 generates no-support transition information 1501 indicating the time-series transition of the health state from the health transition information 701 when the health support is not performed. Hereinafter, a description will be given with reference to FIG.

  FIG. 8 is an explanatory diagram illustrating an example of the transition information 1501 without support according to the first embodiment.

  The transition information without support (health transition without intervention) 1501 indicates the health information of individuals who are not providing health support in time series. The transition information 1501 without support includes health support 1502, elapsed years 1503, personal ID 1504, year 1505, BMI 1506, fasting blood glucose 1507, and information retention years 1508.

  Health support 1502, elapsed years 1503, personal ID 1504, year 1505, BMI 1506, and fasting blood glucose 1507 are health support 706, personal ID 702, year 703, BMI 704, and fasting blood glucose in the health transition information 701 (FIG. 6). Information extracted from 705.

  The elapsed years 1503 indicates the elapsed years from the predetermined reference year to each entry.

  First, in step 105, the health transition information generation unit 1307 extracts information from which health support has not started from the health transition information 701. Specifically, the health transition information generation unit 1307 extracts at least three entries in which the post-health support age 707 of the health transition information 701 is “−” as information that has not started health support. Here, the health transition information generation unit 1307 may extract at least three entries having the same personal ID 702 and a post-health support age 707 of “−”.

  Then, the health transition information generation unit 1307 assigns “none” as the health support 1502 to all the extracted entries. Next, the health transition information generation unit 1307 calculates an elapsed age 1503 for each individual ID 702 for the extracted entry.

  When the post-support transition information 1401 is generated, the predetermined reference year is set to the year 703 when the health support is started. Note that the predetermined reference year of the present embodiment may be any timing as long as the health support effect is not generated, or may be a timing immediately before or after the start of health support.

  On the other hand, when the non-support transition information 1501 is generated, since the individual indicated by the individual ID 702 of each entry has not started health support, the health transition information generation unit 1307 sets the year 703 of all entries as a predetermined reference year. Can be determined.

  For example, when the health transition information generation unit 1307 extracts the health transition information 701 of one individual for three years in the fiscal year 703 in the fiscal year 2000, the fiscal year 2001, and the fiscal year 2002 as information that has not started health support, First, 2000 is set as a predetermined reference year. Then, the health transition information generation unit 1307 calculates 0 year as the elapsed year 1503 of the entry for the year 2000, calculates 1 year as the elapsed year 1503 of the entry for the year 2001, and the elapsed year 1503 of the entry for the year 2002. As a result, 2 years is calculated.

  Furthermore, the health transition information generation unit 1307 determines the year 2001 as a predetermined reference year. Then, the health transition information generation unit 1307 calculates 0 years as the number of years of entry 1503 in 2001. In addition, the health transition information generation unit 1307 calculates one year as the number of years of entry 1503 in 2002.

  Furthermore, the health transition information generation unit 1307 determines the year 2002 as a predetermined reference year. Then, the health transition information generation unit 1307 calculates the year 0 as the elapsed years 1503 of entries for the year 2002.

  By this processing, the health transition information generation unit 1307 generates one entry of the transition information 1501 without support in which the year 1505 indicates the year 2000 and the elapsed year 1503 is 0. Then, the health transition information generation unit 1307 generates two entries of the transition information 1501 without support in which the year 1505 indicates the year 2001, and sets “0” and “1” to the elapsed years 1503 of the generated two entries, respectively. Store.

  In addition, the health transition information generation unit 1307 generates three entries of the transition information 1501 without support in which the year 1505 indicates the year 2002, and “0”, “1”, and “1” are added to the elapsed years 1503 of the generated three entries, respectively. 2 "is stored.

  Then, the health transition information generation unit 1307 calculates the information retention year 1508 of the generated entry. The information retention year 1508 indicates the number of years from the predetermined reference year for each entry to the entry for the latest year 1505 among the entries for each individual ID.

  For example, in the transition information 1501 without support shown in FIG. 8, the health transition information generation unit 1307 has an entry (entry) in which the personal ID 1504 is “0002”, the year 1505 is 2000, and the elapsed year 1503 is “0”. 1509), the entry for the predetermined reference year is also the entry for the year 2000. Further, the no-support transition information 1501 shown in FIG. 8 includes entries for three years in the years 2001 and 2002, including the year 2000. For this reason, the health transition information generation unit 1307 calculates “3” as the information holding year 1508 of the entry 1509.

  Further, an entry (entry 1510) in which the personal ID 1504 is “0002”, the year 1505 is 2001, and the elapsed year 1503 is “0” is an entry in the predetermined reference year. Further, the no-support transition information 1501 illustrated in FIG. 8 includes entries for two years in the year 2002 including the year 2001. For this reason, the health transition information generation unit 1307 calculates “2” as the information retention year 1508 of the entry 1510.

  Further, an entry (entry 1511) in which the personal ID 1504 is “0002”, the year 1505 is 2001, and the elapsed year 1503 is “1” is an entry in the predetermined reference year. Further, the no-support transition information 1501 shown in FIG. 8 includes entries for three years of 2000, 2001, and 2002. For this reason, the health transition information generation unit 1307 calculates “3” as the information holding year 1508 of the entry 1511.

  The information retention year 1508 is used to extract transition information of a predetermined length period from the transition information 1501 without support. For example, when a health transition for five years is necessary, the processing unit described later can extract a health transition for five years only by extracting an entry whose information retention year 1508 is “5”. The health transition information generation unit 1307 stores the generated no-support transition information 1501 in the health transition information storage unit 1352.

  After step 105, the health transition information generation unit 1307 ends the process shown in FIG.

  Next, the process of the health transition prediction unit 1308 will be described. The health transition prediction unit 1308 uses the post-support transition information 1401 and the no-support transition information 1501 stored in the health transition information storage unit 1352, and when the health transition is not performed and the health support is not performed. Predict the health transition of each.

  FIG. 9 is a flowchart illustrating processing by the health transition prediction unit 1308 according to the first embodiment. Hereinafter, processing of each step will be described.

  Based on the instruction input by the input unit 1302, the health transition prediction unit 1308 displays the health information of the candidate (candidate) of the target person who should perform health support, the diagnostic information 301 and the lifestyle information of the health information storage unit 1351. The candidate's health information acquired from 401 is stored in the memory 1305 or the storage medium 1306 (201).

  Note that the data analysis unit 1301 obtains health information at least once from the candidate of this embodiment. The candidate of the present embodiment may be an individual who has provided health support in the past. The health information acquired from the candidate is stored in the diagnosis information 301 and the lifestyle information 401 of the health information storage unit 1351 after the personal ID is assigned. The data analysis unit 1301 predicts the transition of the candidate's health information through subsequent processing.

  The instruction input here may be any instruction as long as it can identify the health information of the candidate who predicts the health transition, and an individual ID for identifying the candidate (personal ID 302 or personal ID 402) for predicting the health transition. It may be included, or a range of personal IDs may be included. Moreover, you may include the year (acquisition year 303 or the reply year 407) when the candidate's health information was acquired.

  After step 201, the health transition prediction unit 1308 acquires the post-support transition information 1401 and the non-support transition information 1501 stored in the health transition information storage unit 1352 and stores them in the memory 1305 or the storage medium 1306 (202).

  After step 202, the health transition prediction unit 1308 selects one health support for predicting the health transition according to the instruction input by the input unit 1302 (203). The instruction input here may be any instruction as long as it indicates the health support to be selected, and may include a name or identifier indicating the health support.

  After step 203, the health transition prediction unit 1308 predicts the transition of the candidate's health information from each of the post-support transition information 1401 and the no-support transition information 1501 based on the candidate's health information acquired in step 201. The transition of the health information used for the purpose is narrowed down (204).

  The effect of health support is generally considered to change depending on the health information (diagnosis information and lifestyle information) of the individual who provides health support. For example, when improving exercise habits as health support, the effect is considered to change according to past exercise habits by individuals.

  Accordingly, in step 204, the health transition prediction unit 1308 narrows down only health information having features similar to the candidate's health information from the post-support transition information 1401 and the non-support transition information 1501.

  More specifically, the health transition prediction unit 1308 changes after support based on health information (health information with intervention) when health support is started and health information (health information without intervention) in a predetermined reference year. The personal ID of the information extracted from the information 1401 and the no-support transition information 1501 is narrowed down.

  For this reason, the health transition prediction unit 1308 is an entry that has a post-health support age of 1403 of “0” and that includes health information similar to the candidate's health information when starting health support. By specifying the entry including health information, the personal ID of the transition of health information extracted from the post-support transition information 1401 is narrowed down.

  In addition, the health transition prediction unit 1308 specifies an entry that has health information similar to the candidate's health information as the health information of a predetermined reference year that is an entry having an elapsed age 1503 of “0”. Thus, the personal ID of the transition of the health information extracted from the transition information 1501 without support is narrowed down. Hereinafter, an example of the narrowing-down method will be described.

  An example of the first narrowing process will be described.

  First, the health transition prediction unit 1308 identifies an entry whose post-health support elapsed time 1403 in the post-support transition information 1401 is “0” and an entry whose elapsed time 1503 in the non-support transition information 1501 is “0”. .

  The health transition prediction unit 1308 compares the information corresponding to the diagnostic information 301 and lifestyle information 401, which are health information included in the identified entry, and the similarity between the candidate's health information (diagnostic information 301 and lifestyle information 401). Is calculated. Then, the health transition prediction unit 1308 compares the calculated similarity with a predetermined threshold, and as a result of the comparison, the personal ID (personal ID 1404) of the entry determined to indicate that the similarity is more similar than the predetermined reference. , And personal ID 1504) as the personal ID of the entry to be extracted.

  In order to calculate the similarity, the health transition prediction unit 1308 generates, for example, a space in which each item of the diagnostic information 301 and the lifestyle information 401 corresponds to one dimension. The health transition prediction unit 1308 then maps the value of each item of the entry specified in the post-support transition information 1401 and the non-support transition information 1501 to the generated space, and the value of each indicator of the candidate's health information The Euclidean distance, the Mahalanobis distance, the inner product, the value of the kernel function defined according to the data, or the like between the points mapped to the generated space is calculated. The health transition prediction unit 1308 uses these calculated values as an index indicating the degree of similarity.

  At this time, the health transition prediction unit 1308 may determine an item (that is, an index for calculating similarity) used for narrowing down according to the health support selected in Step 203. For example, when the effect of the selected health support is set in advance so that a difference occurs according to the current weight and exercise habits of the candidate, the health transition prediction unit 1308 is related to two indicators of weight and exercise habits. Similarity may be calculated to narrow down health information.

  An example of the second narrowing process will be described.

  The health transition prediction unit 1308 obtains the similarity based on the health state definition information 601 defined in advance, and the health information of the individual with the same or closest health state to which the candidate belongs is used as the post-support transition information. It narrows down from each of 1401 and transition information 1501 without support. As the definition of the health state, for example, a definition in which the diagnosis information 301 and the lifestyle information 401 are hierarchized according to the values can be used.

  FIG. 10 is an explanatory diagram illustrating the health condition definition information 601 according to the first embodiment.

  The health condition definition information 601 may be stored in the health condition definition storage unit 1356 of the database 1350. The health condition definition information 601 includes a health condition 602 and an index for defining the health condition (BMI 603, fasting blood glucose 604, blood pressure 605, etc.).

  The index that defines the health state is the same as the index included in the diagnosis information 301 and the lifestyle information 401. For this reason, the index defining the health state may include an index corresponding to the cigarette 404 or the like.

  The health state 602 is a health state determined by the value of an index that defines the health state. In general, a reference threshold is set for the test result, and when the test result exceeds the reference threshold, it is considered that the risk of morbidity increases.

  Therefore, for example, the operator of this embodiment unifies the individual's health condition by defining in advance a health condition determination method that increases the value when the test result of a certain item exceeds a predetermined standard. It is possible to express with a typical index.

  BMI 603, fasting blood glucose 604, and blood pressure 605 indicate the values or ranges of the individual's BMI, fasting blood glucose, and blood pressure when in the health state indicated by health state 602. Therefore, the health transition prediction unit 1308 identifies entries in which the BMI 603, fasting blood glucose 604, and blood pressure 605 correspond to the individual BMI 305, fasting blood glucose 306, and blood pressure, and the health state 602 of the identified entry. Is determined according to the health condition of the individual.

  For example, when the health condition definition information 601 shown in FIG. 10 is defined, when the candidate's BMI is 24.9 or less, fasting blood glucose is 125 or less, and blood pressure is 139 or less, health transition prediction The unit 1308 determines 1 as the candidate's health status.

  The health condition definition information 601 includes, as a health condition determination method, for example, an index exceeding a predetermined threshold regarding the index of diagnostic information 301 and lifestyle habit information 401 that is an index that defines an individual's health condition, Alternatively, the number of indexes that are predetermined values may be totaled, and the total value may be used as a value indicating a health condition.

  The process in step 204 using the health condition definition information 601 will be described. First, the health transition prediction unit 1308 determines a candidate's health state using the candidate's diagnosis information 301 and lifestyle information 401.

  Next, the health transition prediction unit 1308 identifies an entry whose post-health support age information 1401 of the post-support transition information 1401 is “0” and an entry whose post-health transition information 1501 has an elapsed year 1503 of “0”. To do. Then, the health transition prediction unit 1308 determines a health state for each identified entry.

  Then, the health transition prediction unit 1308 obtains the difference between the health status of each entry of the post-support transition information 1401 and the non-support transition information 1501 and the health status of the candidate as the similarity. Then, the health transition prediction unit 1308 uses the personal IDs (individual ID 1404 and personal ID 1504) of the entries indicating that the obtained similarity is the same or closest as the personal ID of the transition information to be narrowed down after the support. 1401 and no-support transition information 1501 are specified. Further, the personal IDs of entries whose similarity is within a predetermined range may be specified.

  By using the health condition definition information 601, the personal ID is narrowed down according to the meaning (health condition) indicated by the health information, so that the individual transition information that is substantially similar to the candidate's status (supported transition information 1401 and no support) Transition information 1501) can be identified and used for prediction.

  Here, the health transition prediction unit 1308 may use the health state definition according to the health support selected in step 203. For example, when the health transition prediction unit 1308 is set in advance that the effect of the selected health support largely depends on the fasting blood glucose of the candidate, and wants to evaluate the fasting blood glucose of the candidate in a finer hierarchy The personal IDs may be narrowed down using the health state definition information 601 that is finely layered with respect to fasting blood glucose.

  After narrowing down the personal ID of the entry to be extracted by using any one of the above-described narrowing methods, the health transition prediction unit 1308 includes the personal ID narrowed down from the post-support transition information 1401 and the non-support transition information 1501. By extracting the entries, it is possible to narrow down the health information used for prediction.

  The above is the process of step 204.

  In step 204, the health transition prediction unit 1308 determines that the personal information ID 1408 of the post-support transition information 1401 is 3 years or more and the personal ID 1508 of the non-support transition information 1501 is 3 years or more. You may narrow down the information only. This is because the health transition prediction unit 1308 predicts health information at three or more timings in the process of predicting health transition described later.

  After step 204, the health transition prediction unit 1308 predicts the health transition after the health support is performed and the health transition when the health support is not performed based on the information narrowed down in step 204 (205). The health transition prediction unit 1308 statistically aggregates each of the transition information after support and the transition information without support included in the information used for the prediction for each health transition year or each transition year (support effect prediction transition 801), Calculated as a predicted value.

  The process of step 205 will be described with reference to FIGS. 11 and 12 below.

  FIG. 11 is an explanatory diagram illustrating an example of a support effect prediction transition 801 according to the first embodiment.

  The support effect prediction transition (predicted transition with intervention) 801 includes a value predicted based on the information narrowed down from the post-support transition information 1401 in step 204. Also, the support effect prediction transition 801 indicates the transition of the predicted value of the candidate's health information at a plurality of timings after the candidate has temporarily provided health support. The support effect prediction transition 801 includes a personal ID 802, health support 803, elapsed years after health support 804, and health information (BMI 805, fasting blood glucose 806).

  The entry of the support effect prediction transition 801 indicates a statistical value of health information obtained according to the number of years that have passed since the health support was performed.

  The personal ID 802 is a personal ID that identifies a candidate who predicts the transition of health information. The health transition prediction unit 1308 stores the personal ID of the candidate in the personal ID 802 in step 205.

  The health support 803 is information such as a name or an identifier indicating health support, and stores information stored in the health support 1402 of the post-support transition information 1401.

  In step 205, the health transition prediction unit 1308 calculates the statistical value of the health information of the entries narrowed down from the post-support transition information 1401 for each health support, each elapsed year after the health support, and each index of the health information. The health transition prediction unit 1308 stores the calculated statistical value in an entry in which health support information is stored in the health support 803.

  The number of years since health support 804 indicates the timing at which the health information used to calculate the statistical value indicated by the entry of the support effect prediction transition 801 is acquired, using the number of years that have passed since the year of health support. .

  The health information (BMI 805, fasting blood glucose 806) is a statistical value calculated for each hierarchy and for each index of health information by collecting post-support transition information 1401 into entries (hierarchies) for each number of years since health support. The health information shown in FIG. 11 shows statistical values of an average value of BMI and a 95% confidence interval, and an average value of fasting blood glucose and a 95% confidence interval.

  In step 205, the health transition prediction unit 1308 divides the narrowed entries of the transition information after support 1401 for each health support (health support 1402) and for each number of years since health support (years after health support 1403). A statistical value is calculated for each index of the divided entry group. The health transition prediction unit 1308 stores the calculated statistical value in the health information of the support effect prediction transition 801 as a predicted value of the transition of the candidate's health information.

  For example, the support effect prediction transition 801 shown in FIG. 11 indicates that, when health support “intervention A” is performed on a candidate whose personal ID 802 is “1001”, when one year has passed since the health support has been performed, The predicted value is 27, and the 95% confidence interval width of the predicted value is “1”.

  The health transition prediction unit 1308 described above calculates an average value and a 95% confidence interval width as statistical values stored in the health information of the support effect prediction transition 801, but the median, deviation, mode, and 95 Statistical values such as confidence interval widths other than% may be calculated. The health transition prediction unit 1308 can express a more accurate predicted value by storing a statistical value in the health information of the support effect prediction transition 801.

  The health transition prediction unit 1308 stores the support effect prediction transition 801 including the predicted value in the prediction information storage unit 1353.

  FIG. 12 is an explanatory diagram illustrating an example of a prediction transition 1601 without support predicted in the first embodiment.

  Predicted transition without support (predicted transition without intervention) 1601 is generated based on the information narrowed down from transition information without support 1501 in step 204. The no-support prediction transition 1601 includes a personal ID 1602, an elapsed year 1603, and health information (BMI 1604, fasting blood glucose 1605).

  The personal ID 1602 is an identifier (personal ID) that identifies a candidate who predicts health transition. The health transition prediction unit 1308 stores the personal ID of the candidate in the personal ID 1602.

  The elapsed year 1603 indicates the time when the health information and the like used to calculate the statistical information indicated by the entry of the no-support prediction transition 1601 is acquired using the number of years that have passed since a predetermined reference year.

  Health information (BMI 1604, fasting blood glucose 1605) is statistics of the average value and confidence interval of BMI and fasting blood glucose calculated for each hierarchy by collecting the transition information without support included in the narrowed down information for each elapsed year. Indicates the value.

  The health transition prediction unit 1308 calculates the statistical values of the entries narrowed down from the no-support transition information 1501 for each elapsed year, and the calculated statistical values are predicted values for the entries in which “none” is stored in the health support 803. Store as. The health transition prediction unit 1308 can express a more accurate predicted value by storing a statistical value in the health information of the transition information 1501 without support.

  For example, the no-support prediction transition 1601 shown in FIG. 12 is a predicted value of the candidate's BMI when one year has elapsed from the reference year when health support is not performed for the candidate whose personal ID 802 is “1001”. Is 28, indicating that its 95% confidence interval width is 2. The health transition prediction unit 1308 calculates the average value and the confidence interval width as statistical values stored in the health information, but may calculate statistical values such as a median value and a mode value.

  In step 205, the health transition prediction unit 1308 stores the generated support effect prediction transition 801 and the no-support prediction transition 1601 in the prediction information storage unit 1353.

  Next, the process of the duration prediction unit 1309 will be described. The duration prediction unit 1309 uses the support effect prediction transition 801 and the no support prediction transition 1601 stored in the prediction information storage unit 1353 to calculate the duration of the effect of health support to be implemented from now on.

  FIG. 13 is a flowchart illustrating processing performed by the duration prediction unit 1309 according to the first embodiment. Hereinafter, processing of each step will be described.

  First, the duration prediction unit 1309 displays a candidate's support effect prediction transition 801 and no support prediction transition 1601 for predicting the duration of the health support effect in accordance with an instruction input from the input unit 1302, and a prediction information storage unit 1353. And stored in the memory 1305 or the storage medium 1306 (1701). The instruction input here may be a personal ID indicating a candidate or a range of personal IDs.

  In step 1701, the duration prediction unit 1309 determines an index of health information for calculating the duration of the health support effect based on the instruction input from the input unit 1302 (1702). The instruction input here indicates indexes of diagnosis information and lifestyle information included in the support effect prediction transition 801 and the no support prediction transition 1601, and indicates BMI, fasting blood glucose, tobacco, and the like.

  After step 1702, the duration prediction unit 1309 predicts the duration of the health support effect using the support effect prediction transition 801 or the support effect prediction transition 801 and the no support prediction transition 1601 acquired in step 1701. (1703).

  FIG. 14 is an explanatory diagram illustrating a method for predicting an effect duration according to the first embodiment.

  FIG. 14 is a diagram in which the obtained support effect prediction transition 801 and no support prediction transition 1601 are plotted as a two-dimensional graph. Here, a case where BMI is determined as an index for predicting the duration in step 1702 is shown as an example.

  An axis 901 shown in FIG. 14 is an axis indicating the years 804 and 1603 after health support, and the larger the value, the more years have passed since health support (or after a predetermined reference year). The axis 902 is an axis indicating the statistical value for each year of the index determined in step 1702.

  A point 903 is the current BMI value of the candidate, and is the BMI value before the health support is performed. The duration prediction unit 1309 plots the BMI value of the candidate's health information acquired in step 201 as a point 903.

  Points I (1), I (2), and I (3) indicate predicted values of health information when health support is implemented, and the number of years since health support (years after health support 804) is Statistical values predicted for BMI when they are 1 year, 2 years and 3 years, respectively. A solid line 904 connecting the point 903, the point I (1), the point I (2), and the point I (3) is a line that represents the transition of the BMI when the health support is performed by linear interpolation.

  Points N (1), N (2), and N (3) indicate predicted values of health information when health support is not performed, and the elapsed years (elapsed years 1603) are 1 year and 2 years, respectively. The predicted statistics for BMI at 3 years are shown. A one-dot chain line 905 that connects the point 903, the point N (1), the point N (2), and the point N (3) is a line that represents the transition of the BMI when the health support is not performed by linear interpolation. .

  A width 906 indicates a confidence interval width of the predicted value. Hereinafter, a method for calculating the duration of the health support effect will be described.

  In step 1703, the duration prediction unit 1309 is the same as the value of the point 903 or worse than the BMI of the point 903 compared to the BMI before the health support (point 903) (becomes a larger value). ) The period up to the BMI year is calculated as the duration of the health support effect. In FIG. 14, the BMI indicated by the point I (3) is the same value as the BMI at the point 903. For this reason, the duration prediction unit 1309 calculates 3 years, which is the range of the period of the line segment 907, as the duration. Thereby, the duration prediction unit 1309 can predict the duration.

  The duration prediction unit 1309 may calculate the duration until the difference between the predicted value of the support effect prediction transition 801 and the predicted value of the no support prediction transition 1601 falls within a predetermined range as the duration. By calculating the duration using the predicted value when the health support is not performed, the duration can be accurately predicted according to the change in the individual's health condition.

  In step 1703, the duration prediction unit 1309 summarizes the calculated durations for each individual ID and health support of the candidate, and stores the durations in the duration storage unit 1355 as duration information 1001. Thereafter, the duration prediction unit 1309 ends the process shown in FIG.

  In the above description, the duration is calculated using diagnostic information such as BMI. However, the duration prediction unit 1309 may calculate the duration using lifestyle information such as drinking. Specifically, the duration prediction unit 1309 calculates the duration while the average daily drinking after starting health support is lower than the average daily drinking at the time of starting health support. Also good.

  FIG. 15 is an explanatory diagram illustrating an example of the duration information 1001 according to the first embodiment.

  The duration information 1001 includes a personal ID 1002, health support 1003, duration prediction 1004, and a calculation index 1005.

  The personal ID 1002 is an identifier indicating a candidate, and is the same identifier as the personal ID 802 of the support effect prediction transition 801 and the personal ID 1602 of the no support prediction transition 1601.

  The health support 1003 is an identifier for identifying the health support whose effect has been predicted, and uses the same name or identifier as the health support 803 (FIG. 11) of the support effect prediction transition 801.

  The duration prediction 1004 indicates the calculated effect duration.

  The calculated index 1005 indicates the index of health information used for calculating the value of the duration prediction 1004, and indicates the index determined in step 1703 shown in FIG.

  Next, processing of the support target person selecting unit 1310 will be described. Based on the duration information 1001 generated by the duration prediction unit 1309 and stored in the duration storage unit 1355, the support target person selection unit 1310 selects a candidate who should perform the health support for each health support as a candidate. Choose from.

  FIG. 16 is a flowchart illustrating processing performed by the support target person selecting unit 1310 according to the first embodiment. Hereinafter, processing of each step will be described.

  The support target person selecting unit 1310 reads the health transition information from the health transition information storage unit 1352 and stores it in the memory 1305 or the storage medium 1306 in accordance with the instruction input by the input unit 1302 (1101).

  The instruction input in step 1101 may be any instruction as long as the health transition information (post-support transition information 1401 and non-support transition information 1501) read from the health transition information storage unit 1352 can be specified.

  After step 1101, the support target person selecting unit 1310 reads from the prediction information storage unit 1353 the predicted transition of the candidate identified according to the instruction input through the input unit 1302. Specifically, the support effect prediction transition 801 indicated by the personal ID 802 and the no-support prediction transition 1601 indicated by the specified personal ID 1602 are read from the prediction information storage unit 1353 and stored in the memory 1305 or the storage medium 1306. Store (1102).

  The instruction input in step 1102 may be any instruction as long as a candidate is specified. The year 703 included in the health transition information 701, diagnostic information (BMI 704, fasting blood glucose 705), lifestyle information, information on health support, etc. Candidates may be indicated according to the conditions. Then, in step 1102, the support target person selecting unit 1310 may read the entry of the personal ID 802 that satisfies the condition from the prediction information storage unit 1353 as the candidate entry.

  In addition, in step 1102, the support target person selecting unit 1310 may determine the health state for each individual indicated by the health transition information 701 using the method used in step 204 and the health state definition information 601. . In step 1102, the support target person selecting unit 1310 may specify an individual whose health condition is worse than a predetermined standard as a candidate, and read the entry of the specified candidate from the prediction information storage unit 1353.

  Further, for example, the support target person selecting unit 1310 receives health support information by receiving conditions such as the presence / absence of past health support, health support performed in the past, or the number of years since the start of health support. Candidates may be identified based on 501.

  After step 1102, the support target person selecting unit 1310 reads out the duration information 1001 indicated by the personal ID 1002 of the identified candidate from the duration storage unit 1355 and stores it in the memory 1305 or the storage medium 1306 (1103).

  After step 1103, the support target person selecting unit 1310 determines health support for determining the target person and conditions for selecting the target person to perform the health support (1104). Specifically, the support target person selecting unit 1310 determines according to the condition input from the input unit 1302. The predetermined conditions to be input are, for example, health support for determining a target person, the number of target persons to be selected for each health support, and the like.

  After step 1104, the support target person selecting unit 1310 selects the target person from the candidates for the health support determined in step 1104 based on the duration prediction 1004 of the duration information 1001 read in step 1103 ( 1105).

  A first example of the selection method in step 1105 will be described. The first example is a method of selecting a target person based on the duration prediction 1004 and a predetermined time standard.

  In step 1105, the support target person selecting unit 1310 compares the values of all the duration predictions 1004 included in the read duration information 1001. Then, the support target person selecting unit 1310 selects a candidate having the largest duration prediction 1004 among candidates not yet determined as a target person, that is, a candidate who can be expected to have the longest effect of health support. Select as a target for health support.

  Further, the support target person selecting unit 1310 may not select a candidate as a target person by excluding candidates for the duration prediction 1004 that is lower than the lower limit value of the predetermined duration.

  Furthermore, a second example of the selection method in step 1105 will be described. The second example is a method of selecting a candidate whose duration prediction 1004 and other information correspond to a predetermined standard as a target person.

  In the second example of step 1105, the support target person selecting unit 1310, the value of the duration prediction 1004 included in the read duration information 1101, the support effect prediction transition 801 and the no support prediction transition 1601 read in step 1102. The target person may be selected from the candidates using the information.

  For example, the support target person selecting unit 1310 has the health information of the entry in which the post-health support elapsed time 804 of the read support effect prediction transition 801 is “1” and the elapsed time 1603 of the no-support prediction transition 1601 is “1”. The difference from the health information of a certain entry is calculated for each individual ID. Accordingly, the support target person selecting unit 1310 can calculate the magnitude of the effect immediately after the health support is performed.

  Then, the support target person selecting unit 1310 selects, as the target person, a candidate whose calculated difference is larger than a predetermined reference (an average value or an intermediate value of the difference) and whose duration prediction 1004 is largest. To do. Accordingly, the support target person selecting unit 1310 selects a candidate using the calculated difference and the two values of the duration prediction 1004, thereby selecting a target person who is most effective and has a large effect. You can choose.

  In step 1105, the support target person selecting unit 1310 determines whether to select each candidate as a target person who should perform health support by selecting a candidate who is a target person based on the duration prediction 1004. it can.

  After step 1105, the support target person selecting unit 1310 determines whether it is necessary to additionally select the target person based on the condition determined in step 1104 (1106). For example, when the upper limit value of the number of target persons is determined in step 1104, the support target person selecting unit 1310 determines whether the number of selected target persons has reached the upper limit value in step 1106, and sets the upper limit value. If not, the process returns to step 1105 to additionally select a target person.

  When the number of selected target persons reaches the upper limit value, the support target person selecting unit 1310 does not additionally select the target person and outputs information (personal ID or the like) regarding the selected target person to the output unit 1303 ( 1107).

  Here, the support target person selecting unit 1310 may output the health transition information read in Step 1101 and the information based on the predicted transition read in Step 1102 as a screen 1800 described later. After step 1107, the support target person selecting unit 1310 ends the process shown in FIG.

  FIG. 17 is an explanatory diagram illustrating an example of a user interface screen 1800 according to the first embodiment.

  The screen example illustrated in FIG. 17 is an example of a screen 1800 displayed on the output unit 1303. FIG. 17 shows a screen 1800 for selecting a target person for health support based on the prediction result (duration information 1001) of the duration of the effect shown in FIG. The screen 1800 includes a region 1801 to a region 1805, a region 1811, and a region 1821.

  A region 1801 is a numerical value input box for inputting the number of subjects to be selected. An area 1801 is an area for inputting conditions used in step 1104 shown in FIG.

  A region 1802 is an input box for selecting health support content for selecting a target person. An area 1802 is an area for inputting conditions used in step 1104 shown in FIG.

  An area 1803 is a button for instructing to read data related to the candidate from the database 1350. When the area 1803 is operated, steps 1101 to 1103 shown in FIG. 16 are executed.

  A region 1804 is an input box for inputting an index for calculating the duration of the health support effect. An area 1804 is an area for inputting an index used in step 1702 shown in FIG.

  An area 1805 is an input box for starting a process of selecting a target person. When the area 1805 is operated, the processing after step 1104 shown in FIG. 16 is executed.

  An area 1811 is an area for displaying information on the selected target person. The region 1811 includes regions 1812 to 1816.

  An area 1812 is a histogram representing the number distribution of each index used when selecting a target person. For example, by specifying health information such as BMI or fasting blood glucose as an index displayed in the area 1812, the support target person selecting unit 1310 displays the number distribution for each value of the health information. An area 1812 shown in FIG. 17 displays the ratio of the number of persons selected as subjects in each distribution using a white histogram and a black histogram.

  An area 1813 is an input box for designating an index for displaying the number distribution in the area 1812. An area 1813 illustrated in FIG. 17 indicates that BMI is designated.

  An area 1814 is an area for displaying the average improvement amount from the start of health support to the next year of the index displayed for each index specified in the area 1813 and used for calculating the duration. The support target person selecting unit 1310 calculates the average improvement amount displayed in the area 1814 based on a difference in each index between the target person's support effect prediction transition 801 and the no support prediction transition 1601.

  An area 1815 is an area that is displayed for each index specified in the area 1813 and indicates an effect duration (standby time). The support target person selecting unit 1310 calculates the effect duration displayed in the region 1815 based on the duration prediction 1004 of the target person duration information 1001.

  An area 1816 is a data table indicating information on the selected target person, and shows an estimated average improvement amount, a range of duration, and the like.

  An area 1817 is a table showing information on the selected target person, and displays the personal ID, age, duration, and the like of the selected target person.

  An area 1818 is a display indicating the individual currently selected by the operator or the like in the area 1817 in order to confirm the detailed information of the individual from the selected target person. The operator or the like arbitrarily selects the target person in the area 1817 via the input unit 1302.

  An area 1821 shows the detailed information of the individual of the subject selected in the area 1817. An area 1821 displays details of effect transition prediction after the health support is performed. Information on the individual of the subject specified by the area 1818, information on effect transition prediction, and health transition prediction are displayed. The region 1821 includes a region 1822 to a region 1825.

  An area 1822 is a screen that displays detailed information of the subject currently selected in the area 1817. In addition to age (age 304) and duration of effect (duration prediction 1004), diagnosis information 301 or lifestyle information Information included in 401 is displayed.

  An area 1823 is a two-dimensional graph showing a time-series transition of the effect duration calculation index used for calculation of the effect duration (step 1703), and corresponds to the screen example shown in FIG. An area 1823 displays respective time series transitions with and without support.

  An area 1824 is an input box for designating an index for displaying a time-series transition related to an index other than the calculation index for the duration in the area 1825. A region 1825 is a two-dimensional graph showing the transition of the index specified in the region 1824, and displays respective time series transitions with and without support.

  The operator, etc. can grasp the result that can be expected when the target person is provided with health support by referring to the screen 1800 shown in FIG. 17, or the information indicating the target person such as the age of the target person. can do.

  Hereinafter, modifications of the first embodiment will be described.

  In the first embodiment, as examples of the health condition definition information 601 stored in the health condition definition storage unit 1356, candidates are hierarchically based on the values of the diagnosis information 301 and the lifestyle information 401 between individuals and candidates. An example was given to define the health status of the elderly. However, the definition method of the health condition definition information 601 is not limited to this.

  For example, the health condition definition information 601 may be defined by an expression having an index included in the diagnosis information 301 and the lifestyle information 401 as an explanatory variable. Specifically, the health condition definition information 601 sets the index X calculated as “X = K1 × BMI + K2 × fasting blood glucose level + K3 (K1, K2, and K3 are predetermined real numbers)” as the health condition. An expression may be defined. In this case, the health condition definition information 601 may be represented not by a table format as shown in FIG. 10 but by a collection of at least one expression.

  Furthermore, the health condition definition information 601 may directly use the value of one index included in the diagnostic information 301 or the lifestyle information 401. For example, the health condition definition information 601 may hold a BMI value or a fasting blood glucose level as a health condition definition.

  Further, in the above-described first embodiment, the example in which BMI is used as the calculation index of the duration in the calculation process of the effect duration shown in FIG. 13 by the duration prediction unit 1309 is shown. However, the duration prediction unit 1309 may use an index other than BMI as a calculation index for the duration.

  Specifically, the duration prediction unit 1309 may use another index included in the support effect prediction transition 801, for example, fasting blood glucose level or lifestyle information as an index for calculating the duration.

  Further, the duration prediction unit 1309 may use an index obtained by combining a plurality of indices. Specifically, the duration prediction unit 1309 may be defined by an expression having an index included in the diagnosis information 301 and the lifestyle information 401 as an explanatory variable.

  As an example, the duration prediction unit 1309 calculates the effect duration using the index X calculated as “X = K1 × BMI + K2 × Fasting blood glucose level + K3 (K1, K2, and K3 are predetermined real numbers)”. It may be used as an indicator. Further, the duration prediction unit 1309 may use a different index for calculating the effect duration for each health support.

  Further, in the above-described first embodiment, in the duration calculation processing (step 1703 in FIG. 13) by the duration prediction unit 1309, the value is the same as the value at the start of health support compared to the index value at the start of health support. Alternatively, an example is shown in which the period until the year of a value worse than the index value at the start of health support is calculated as the effect duration. However, the duration calculation method by the duration prediction unit 1309 is not limited to this.

  Specifically, the duration prediction unit 1309 varies the duration based on the prediction result of the index value transition after the health support is implemented and the prediction result of the index value transition when the health support is not implemented. It may be calculated by various methods. Hereinafter, a case where BMI is used as an index for calculating the duration will be described as an example.

  A calculation example of the first duration will be described.

  The duration prediction unit 1309 may determine a threshold value from the value of the index at the start of health support and the value of the index after the implementation of health support, and calculate the duration. For example, when the value of BMI at the start of health support is 25 and the value of BMI in the first year after the health support is 21, the duration prediction unit 1309 sets a threshold T1 for calculating the duration. , T1 = K × 25 + (1−K) × 21 (K is an arbitrary real number).

  Then, the duration prediction unit 1309 may calculate the duration until the BMI value exceeds the threshold T1 as the duration. For example, the duration prediction unit 1309 sets K = 0.5. Accordingly, the duration prediction unit 1309 may calculate the duration until the improvement amount returns to the amount of 50% as the effect duration when the BMI value improved by the health support is 100%.

  By setting the threshold value T1 in this way, the duration prediction unit 1309 can calculate the duration of health support based on the amount of improvement effect by health support.

  A calculation example of the second duration will be described.

  The duration prediction unit 1309 calculates the duration using the prediction result of the transition of the index value after health support and the prediction result of the transition of the index value when the health support is not performed.

  For example, the duration prediction unit 1309 calculates the difference D = N−I between the predicted value I of BMI after the health support is performed and the predicted value N of BMI when the health support is not performed, and the difference D and a predetermined value The threshold is compared every year. Then, the duration prediction unit 1309 may calculate the duration until the year when the difference D is smaller than a predetermined threshold as the duration.

  The difference D indicates the amount of the effect of health support calculated for each predicted period. For example, the BMI after one year of health support is 21 and the BMI after two years is predicted to be 23, and the BMI after one year without health support is 25. Assume that the later BMI was predicted to be 26. At this time, the difference D (effect) with respect to BMI one year after the start of health support is 25-21 = 4, and the difference D (effect) two years after the start of health support is 26-23 = 3.

  Here, when the predetermined threshold value for the difference D is 3 or less, the condition indicated by the predetermined threshold value is satisfied two years after the start of health support. For this reason, the duration prediction unit 1309 calculates 2 years as the duration.

  In this way, the duration is predicted by calculating the duration using the prediction result of the transition of the index value after the implementation of health support and the prediction result of the transition of the index value when the health support is not implemented. The unit 1309 can calculate the duration of the effect more accurately because it can specify the time when the effect is reduced based on the prediction result compared with the case where the health support is not actually performed.

  A calculation example of the third duration will be described.

  In the calculation of the effect duration, the duration prediction unit 1309 includes a 95% confidence interval width of the index of the diagnostic information and lifestyle information (in this case, the predicted value) indicated by the support effect prediction transition 801 and the no support prediction transition 1601 , Information on the variation of values is used.

  For example, the duration prediction unit 1309 identifies the lower end or upper end value of the BMI range based on the predicted BMI value and the confidence interval width when there is health support, A predicted value of BMI in the case of no support may be compared and the duration may be calculated based on the comparison result.

  For example, if the duration prediction unit 1309 specifies the range of the predicted value of BMI when health support is provided as 21 to 25 and the predicted value of BMI without health support is 26, the prediction of BMI without health support is performed. The value is larger than the upper limit value 25 of BMI with health support. For this reason, with the probability of 95% or more, it can be said that the effect of health support is still sustained.

  When the predicted value of BMI without health support is 20, the predicted value of BMI without health support is smaller than the lower limit value 21 of BMI with health support. For this reason, it can be said that the effect of health support is not sustained with a probability of 95% or more. For this reason, the duration prediction unit 1309 may calculate the duration until the year when the predicted value of BMI without health support falls below the lower limit value of BMI with health support as the duration.

  In this way, by calculating the duration using the confidence interval width, the duration prediction unit 1309 can calculate the duration in consideration of the variation in the prediction result, and can improve the reliability of the duration.

  As described above, the analysis system according to the first embodiment predicts the transition of health information after the health support is performed based on the health information and the history indicating the health support performed in the past, and the effect of the health support. Predict the duration of. And since the target person who implements health support is selected based on the predicted duration, it is possible to select a more appropriate target person by implementing health support.

  In the first embodiment, an analysis system that predicts a duration of the effect of health support based on health information and a history indicating health support performed in the past, and selects a target person for health support based on the predicted duration. Explained the example. In Example 2, an example of an analysis system that formulates a long-term support plan for each target person based on the predicted duration will be described. Here, the long-term support plan is a plan for providing health support multiple times to a target person.

  The configuration of the analysis system of the second embodiment is the same as the configuration shown in FIG. In addition, the health transition information generation unit 1307, the health transition prediction unit 1308, the duration prediction unit 1309, and the support target person selection unit 1310 of the second embodiment execute the same processing as that of the first embodiment.

  However, in the second embodiment, the long-term support plan formulation unit 1312 executes the process shown in FIG.

  FIG. 18 is a flowchart illustrating processing by the long-term support plan formulation unit 1312 according to the second embodiment.

  Before the long-term support plan formulation unit 1312 executes the process shown in FIG. 18, the target person who performs health support and the health support to be performed for the first time are already selected. The target person and the first health support may be determined by the process shown in FIG. 2 or may be instructed via the input unit 1302.

  In particular, when the process shown in FIG. 18 is executed, the process shown in FIG. 5 has been executed, and the health transition information storage unit 1352 holds post-support transition information 1401 and no-support transition information 1501. The following process shows a process when a long-term support plan is formulated when a predetermined target person implements health support. When there are a plurality of target persons, the long-term support plan formulation unit 1312 A plan is formulated by repeating the process shown in FIG. 18 for each target person.

  The long-term support plan formulation unit 1312 acquires and determines a plan period for formulating a long-term support plan in accordance with the instruction input to the input unit 1302 (1201). For example, when formulating an implementation plan for health support for four years from now, the long-term support plan formulation unit 1312 determines four years as the planning period.

  Here, the input instruction may include any information as long as the planning period can be specified. For example, the input instruction may include a start date and an end date of the planning period, or may include only the period of the planning period.

  After step 2201, the long-term support plan formulation unit 1312 determines whether the plan period determined in step 1201 is longer than the duration prediction 1004 of the first health support performed by the subject (1204).

  The long-term support plan formulation unit 1312 determines that it is not necessary to plan health support for the second time or later when the plan period is the duration prediction 1004 or less. Therefore, the long-term support plan formulation unit 1312 determines information indicating only the currently planned health support as a long-term support plan (1206), and ends the processing shown in FIG.

  For example, when the long-term support plan formulation unit 1312 determines the plan period for formulating the plan as 4 years and the duration prediction 1004 is 5 years, the plan period is less than the duration prediction 1004, and within the plan period, The effect of health support can be expected to last. For this reason, the long-term support plan formulation unit 1312 again determines that it is not necessary to implement health support, and ends the process illustrated in FIG.

  On the other hand, when the long-term support plan formulation unit 1312 determines the period for formulating the plan as 4 years and the duration prediction 1004 is 2 years, it is considered that it is necessary to implement health support again after 2 years.

  For this reason, when the planning period is longer than the duration prediction 1004, the long-term support plan formulation unit 1312 updates the planning period by subtracting the planning period by the value of the duration prediction 1004. Then, step 1205 is executed to plan health support for the second and subsequent times.

  In step 1204, the long-term support plan formulation unit 1312 stores the value of the duration prediction 1004 and other processing results in the memory 1305 and the storage medium 1306.

  In step 1205, the long-term support plan formulation unit 1312 obtains predicted values of each index of health information after the duration of the first health support from the support effect prediction transition 801 and the no-support prediction transition 1601. Here, the long-term support plan formulation unit 1312 acquires the acquired predicted value as the health information of the subject after the duration of the effect of the first health support.

  That is, the support effect prediction transition 801 and the no-support prediction transition 1601 indicate predicted values of future health information predicted for each subject in the presence or absence of health support. Specifically, values such as diagnostic information (BMI, fasting blood glucose) and lifestyle information include information calculated for each elapsed year. For this reason, in step 1205, the long-term support plan formulation unit 1312 acquires the predicted value of each index of health information corresponding to the number of years elapsed after the duration as health information.

  After step 1205, the health transition prediction unit 1308 acquires the health information acquired in step 1205 instead of the health information acquired in step 201 shown in FIG. Then, the health transition prediction unit 1308 executes the health transition prediction process shown in FIG. 9 based on the acquired health information, the post-support transition information 1401, and the no-support transition information 1501, and the support effect prediction transition 801 and The prediction transition 1601 without support is stored in the prediction information storage unit 1353 (2202).

  Thereby, the health transition prediction unit 1308 shows the transition of the health information after the second health support based on the health information predicted after the duration has elapsed since the first health support. The support effect prediction transition 801 shown can be generated.

  After step 2202, the duration prediction unit 1309 executes the effect duration prediction process shown in FIG. 13 (2203). Accordingly, the duration prediction unit 1309 generates duration information 1001 related to health support that the subject performs for the second time.

  After step 2203, the long-term support plan formulation unit 1312 determines whether the updated plan period is less than or equal to the duration prediction 1004 of the duration information 1001 related to the second health support after step 1204. If the updated plan period is equal to or less than the second duration prediction 1004, the long-term support plan formulation unit 1312 executes Step 1206.

  When the updated plan period is longer than the second duration prediction 1004, the long-term support plan formulation unit 1312 plans the third and subsequent health support by repeating steps 1205, 2202 and 2203.

  In Step 1206 after executing Step 2203, the long-term support plan formulation unit 1312 performs the first health support and the second and subsequent health support whose duration has been predicted in Step 2203 to the subject. Decide on support. Then, based on the duration prediction 1004 predicted and stored in step 2203, a long-term support plan is formulated.

  For example, in step 1201, the period is determined to be five years, the first health support duration prediction 1004 is two years, the second health support duration prediction 1004 is two years, and the third health support When the duration prediction 1004 is 2 years, the long-term support plan formulation unit 1312 performs health support in step 1206 three times in total (current year (0), 2 years later, and 4 years later). Develop a long-term support plan to begin.

  In step 1206, the long-term support plan formulation unit 1312 obtains information on the target person of the health support plan, information on the health support to be implemented, and information on the timing of the implementation based on the long-term support plan formulated. You may output from the output part 1303. FIG.

  In the process of FIG. 18, the second and subsequent health support may be the same as the first health support. Further, after step 2203, the long-term support plan formulation unit 1312 may determine the second and subsequent health support as a health support different from the first health support based on the predicted duration prediction 1004.

  More specifically, in step 2202, the health transition prediction unit 1308 may predict health transitions when a plurality of health supports are implemented as the types of health transitions to be predicted. Further, in the long-term support plan to be finally formulated, the long-term support plan formulation unit 1312 may plan and output a plurality of types of health support services with different health support.

  For example, the long-term support plan formulation unit 1312 uses the health transition predicted for a plurality of health supports to provide priority lifestyle support for the first health support, and for the second health support. A long-term plan may be generated that combines different health support, such as implementing lifestyle support that is milder than the first.

  As described above, the analysis system according to the second embodiment predicts the duration of the effect from health information and information on past health support, and formulates a long-term support plan for each target person based on the predicted duration. To do.

  In the first embodiment, the health transition prediction unit 1308 performs similar processing based on health information indicators that affect the effect of health support and knowledge accumulated in advance in the health transition information narrowing process in step 204. A method for changing the degree calculation method for each selected health support is described. Here, the health information that affects the effect of health support is, for example, an index of diagnostic information or lifestyle information.

  The analysis system of Example 3 generates knowledge about health information that affects the effect based on past results, even when the prior knowledge about health information that affects the effect of health support is not held for each health support. Use to improve the reliability of predictions.

  The configuration of the analysis system of the third embodiment is the same as that shown in FIG. However, the health transition prediction unit 1308 according to the third embodiment executes processing different from the processing according to the first embodiment. In addition, the health index narrowing unit 1311 performs the following processing in the third embodiment.

  First, the process of the health index narrowing unit 1311 will be described. The health index narrowing-down unit 1311 narrows down the index of health information that affects the effect of health support for each health support.

  FIG. 19 is a flowchart illustrating a process performed by the health index narrowing unit 1311 according to the third embodiment.

  The health index narrowing-down unit 1311 determines the health support of the target for narrowing down the health index according to the instruction input to the input unit 1302 (1901). The instruction input here may be any instruction as long as information specifying health support is included.

  After step 1901, the health index narrowing-down unit 1311 acquires post-supporting transition information 1401 and non-supporting transition information 1501 stored in the health transition information storage unit 1352, and stores them in the memory 1305 or the storage medium 1306 (1902). .

  After step 1902, the health index narrowing-down unit 1311 calculates the feature quantity of the health transition for each individual ID based on the health information included in the post-support transition information 1401 and the non-support transition information 1501 (1903). Specifically, the health index narrowing-down unit 1311 calculates the feature quantity of the health transition based on the time-series change of the health information index that is targeted for improvement by the health support determined in step 1901.

  For example, when the determined health support improvement target is BMI, the health index narrowing-down unit 1311 relates to the BMI time-series information from the BMI and age information in the health information included in the no-support transition information 1501. The feature amount is calculated. For example, the feature amount is calculated by applying a known time series modeling method such as polynomial regression or autoregression to the relationship between the BMI and the elapsed years, and obtaining parameters (for example, coefficient values of the polynomial regression equation index). Is calculated as a feature amount. At this time, the health index narrowing unit 1311 may calculate the feature amount using only the post-support transition information 1401.

  After step 1903, the health index narrowing-down unit 1311 narrows down the index of health information that affects the feature value based on the feature value of the health transition calculated for each individual in step 1903.

  The health index narrowing unit 1311 calculates, for example, a value indicating the strength of association between the feature quantity of health transition calculated for each individual ID and each index of health information when the elapsed year is 0, and the feature quantity We narrow down indicators that are strongly related to health information as indicators of health information that affect the effectiveness of health support. The health index narrowing-down unit 1311 can narrow down the index by performing threshold determination on a known index indicating the strength of association, such as Pearson's correlation coefficient or MIC (Maximum Information Coefficient).

  The health index narrowing unit 1311 obtains a value indicating the strength of the relationship between the health transition characteristic amount calculated for each individual ID and each index of the health information at the time of health support implementation in step 1902. Is calculated from the information used to calculate. For example, when the feature amount of the health transition is calculated from the no-support transition information 1501, it is calculated using information for each individual ID included in the no-support transition information 1501. For example, when the feature amount of the health transition is calculated from the post-support transition information 1401, it is calculated using information for each individual ID included in the post-support transition information 1401.

  The health index narrowing unit 1311 stores the narrowed health information index and the information indicating the health support determined in step 1901 in the calculation index storage unit 1357, and ends the process illustrated in FIG.

  The health transition prediction unit 1308 according to the third embodiment stores the health stored in the calculation index storage unit 1357 as an index of health information to be used when calculating the similarity in order to narrow down the health transition information in step 204 shown in FIG. Use information indicators.

  For example, when the calculation index storage unit 1357 stores BMI and fasting blood glucose as indexes strongly related to health support, the health transition prediction unit 1308 of the third embodiment, in step 204, determines the BMI and fasting rate. The similarity is calculated using only blood glucose. Then, the health transition prediction unit 1308 narrows down the transition of health information by specifying the personal ID in the post-support transition information 1401 and the non-support transition information 1501 based on the calculated similarity.

  As a result, the health transition prediction unit 1308 according to the third embodiment can improve the reliability of the prediction of the health transition because the past information used for the prediction can be narrowed down using the health information that affects the health transition. .

  Then, the health index narrowing-down unit 1311 may periodically execute the process shown in FIG. 19 in parallel with the accumulation of health information and health support information 501 which are health support performance data. Thereby, the health transition prediction unit 1308 of Example 3 can perform more accurate prediction based on the latest results.

  Note that the data analysis unit 1301 in the second embodiment may narrow down the index using the health index narrowing-down unit 1311 in the third embodiment. This makes it possible to accurately predict health transitions when formulating a long-term support plan.

  As described above, the analysis system of the third embodiment optimally changes the method for narrowing down health transition information for each health support, even when there is no prior knowledge of health information that affects the effect of health support. Health transitions and duration of health support can be predicted based on actual results.

  The health transition prediction unit 1308 of Example 1 predicted health transition by the process shown in FIG. Specifically, the health transition prediction unit 1308 of the first embodiment extracts health transition information having characteristics similar to the health information of the target person from the past health transition information 701, and statistically calculates them. The process of predicting the health transition of the subject was executed.

  The health transition prediction unit 1308 of Example 4 predicts future health transitions using a graphical model that predicts health transitions constructed from the health transition information 701. The health transition prediction unit 1308 uses a graphical model to extract health transition without extracting health transition information (post-support transition information 1401 and non-support transition information 1501) having characteristics similar to the target person's health transition. It becomes possible to predict. The health transition prediction unit 1308 can execute highly reliable prediction based on all past health transition information.

  The analysis system of Example 4 defines each item of health information and health support information 501 included in health transition information as a random variable, and a graphical model constructed so as to express a stochastic dependency between these random variables. A prediction transition is constructed from predicted values of future health information predicted by inputting the health information of the prediction target person and the target health support information 501 as observation values.

  The configuration of the analysis system of the fourth embodiment is the same as the configuration shown in FIG. In addition, the health transition information generation unit 1307, the duration prediction unit 1309, and the support target person selection unit 1310 of the second embodiment execute the same processing as that of the first embodiment. However, the health transition prediction unit 1308 of the fourth embodiment executes processing different from the processing shown in FIG.

  FIG. 20 is a flowchart illustrating the process of the health transition prediction unit 1308 according to the fourth embodiment.

  Step 201 and step 202 shown in FIG. 20 are the same as step 201 and step 202 shown in FIG.

  After step 202, the health transition prediction unit 1308 predicts the health transition using the graphical model stored in the health transition prediction model storage unit 1358 (2001). This graphical model is a graphical model constructed by using a known method from the health transition information stored in the health transition information storage unit 1352.

  For example, this graphical model is generated using a Bayesian network in which each index of health information of health transition information (post-support transition information 1401 and no-support transition information 1501) is a random variable.

  FIG. 21 is an explanatory diagram schematically illustrating an example of a graphical model according to the fourth embodiment.

  The graphical model shown in FIG. 21 is a model obtained by defining the health information index of the health transition information as a random variable for each elapsed year. When predicting the health transition, the health transition prediction unit 1308 stores the value of the index of the health information of the subject in the value of the random variable whose elapsed years is 0, and based on the information, the random variable for each elapsed year To predict the health transition of the subject.

  For example, when the index value of the current health information of the subject is calculated as the probability variable 2101, the health transition prediction unit 1308 uses the graphical model shown in FIG. 21 to indicate the value shown in the probability variable 2102 after one year. Predict that it will be.

  The graphical model may be set differently for each health support, and the health transition prediction unit 1308 may predict the health transition using the graphical model for each health support, and may relate to the type of health support. The prediction may be performed using a single graphical model including information as a random variable.

  In step 2001, the health transition prediction unit 1308 outputs the predicted health transition as the support effect prediction transition 801 and stores it in the prediction information storage unit 1353. Thereafter, the health transition prediction unit 1308 ends the process shown in FIG.

  As described above, the analysis system according to the fourth embodiment extracts personal health transition information having characteristics similar to the health transition of the target person by predicting health support using a graphical model that predicts health transition. It is possible to predict the health transition without performing, and a highly reliable prediction based on all past health transition information can be performed.

  Note that the data analysis unit 1301 of the second embodiment may use the health transition prediction unit 1308 of the fourth embodiment. In addition, the data analysis unit 1301 according to the third embodiment generates a graphical model for the index narrowed down by the health transition prediction unit 1308 after narrowing down the health information index by the health index narrowing unit 1311 and predicts the health transition. May be.

  The present invention is not limited to the above-described embodiments, and includes various modifications. For example, the above-described embodiments have been described in detail for easy understanding of the present invention, and are not necessarily limited to those having all the configurations described.

  Further, a part of the configuration of one embodiment can be replaced with the configuration of another embodiment, and the configuration of another embodiment can be added to the configuration of one embodiment. In addition, it is possible to add, delete, or replace another configuration for a part of the configuration of each embodiment.

  Each of the above-described configurations, functions, processing units, and the like may be realized by hardware by designing a part or all of them with, for example, an integrated circuit. Also, each of the above-described configurations, functions, and the like may be realized by software by interpreting and executing a program that realizes each function by the processor. Information such as a program, a table, or a file for realizing each function can be placed in a memory, a hard disk, a recording device such as an SSD (Solid State Drive), or a recording medium such as an IC card or an SD card.

  Further, the control lines and the information lines indicate what is considered necessary for the explanation, and not all the control lines and information lines on the product are necessarily shown. In practice, it may be considered that almost all the components are connected to each other.

1301 Analysis system 1302 Input unit 1303 Output unit 1304 Computing device 1305 Memory 1306 Storage medium 1307 Health transition information generation unit 1308 Health transition prediction unit 1309 Duration prediction unit 1310 Support target person selection unit 1311 Health index narrowing unit 1312 Long-term support plan formulation Unit 1350 database 1351 health information storage unit 1352 health transition information storage unit 1353 prediction information storage unit 1354 health support information storage unit 1355 duration storage unit 1356 health state definition storage unit 1357 calculation index storage unit 1358 health transition prediction model storage unit

Claims (15)

An analysis system,
A processor and a storage unit;
The storage unit indicates health information including at least one of an index indicating an individual's health level and an index indicating a factor affecting the health of the individual, and the individual who has performed an intervention supporting health. Store intervention information,
The analysis system includes:
The processor acquires health information of a candidate who performs the intervention, and acquires time-series health information acquired from the start time of the intervention of the individual who has performed the intervention as a health transition with intervention from the storage unit. Based on the candidate's health information and the health transition with intervention, predicting the candidate's time-series health information when the candidate starts the intervention as a prediction transition with intervention, and A health transition prediction unit that stores the predicted transition with intervention in the storage unit;
The processor predicts a duration indicating a period during which the candidate's health is improved as an effect of the intervention using the predicted transition with intervention, and stores the duration in the storage unit. When,
An analysis system, wherein the processor comprises: a subject selection unit that selects a subject to perform the intervention from the candidates based on the duration. The analysis system according to claim 1,
After the duration prediction unit predicts a first duration of the effect of the first intervention and the subject selection unit selects a subject to perform the first intervention, the processor If the plan period during which an intervention is performed is longer than the first duration, the first duration has elapsed since the subject started the first intervention. A long-term planning unit that predicts the health information of the subject at the time based on the predicted transition with intervention,
The health transition prediction unit, after the subject has started a second intervention after the first duration has elapsed, based on the predicted health information of the subject and the health transition with intervention Predicting time-series health information of the subject, and storing the predicted time-series health information in the storage unit,
The duration prediction unit predicts the presence of intervention as a second duration of the effect of the second intervention, a period during which the candidate's health improves when the candidate starts the second intervention Predict using transition,
The long-term planning unit performs the first intervention and the second intervention on the subject when the result of subtracting the first duration from the planning duration is equal to or less than the second duration. An analysis system characterized by deciding to do. The analysis system according to claim 1,
The health transition prediction unit
Obtaining time series health information of individuals who have not performed the intervention from the storage unit as health transition without intervention,
Based on the candidate health information and the health transition without intervention, predicting the candidate time-series health information when the candidate does not perform the intervention as a prediction transition without intervention,
Storing the predicted transition without intervention in the storage unit;
The duration prediction unit predicts the duration using the prediction transition with intervention and the prediction transition without intervention. The analysis system according to claim 3,
The health transition prediction unit
The health information at a predetermined timing when the individual who performed the intervention started the intervention is acquired from the health transition with the intervention as health information with the intervention,
Based on the similarity between the candidate's health information and the health information with intervention, a first individual similar to the candidate is identified from the individuals who have performed the intervention;
Predicting the predicted transition with intervention based on the health transition with intervention of the first individual;
From the health transition without intervention, to obtain the health information of the earliest timing as health information without intervention,
Based on the similarity between the candidate's health information and the health information without intervention, a second individual similar to the candidate is identified from individuals who have not performed the intervention;
An analysis system for predicting the predicted transition without intervention based on the health transition without intervention of the second individual. The analysis system according to claim 4,
The storage unit further stores state definition information for determining the health state of the individual from the health information,
The health transition prediction unit
Based on the state definition information, a health state determined from the health information with the intervention, a health state determined from the health information without the intervention, and a health state determined from the health information of the candidate are specified. ,
Identifying the first individual based on the similarity between the health condition determined from the candidate's health information and the health condition determined from the identified health information with intervention;
The analysis system characterized in that the second individual is specified based on a similarity between a health condition determined from the candidate's health information and a health condition determined from the specified health information without intervention. The analysis system according to claim 4,
The health information includes a plurality of the indicators,
Based on the health transition information with the intervention, having a health index narrowing unit that identifies the index of the health information that affects the effect of the intervention,
The health transition prediction unit
In the identified indicator, the first individual is identified based on the similarity between the candidate's health information and the health information with intervention,
The analysis system characterized in that the second individual is specified based on a similarity between the health information of the candidate and the health information without intervention in the specified index. The analysis system according to claim 1,
The health transition prediction unit
The health information of the time series of the start and after the start of the intervention of a plurality of individuals who have performed the intervention is acquired from the storage unit as health transition with the intervention of the plurality of individuals,
Predicting the predictive transition with intervention by calculating statistical values of health information for each of a plurality of timings included in the health transition with intervention of the plurality of individuals,
The analysis system is characterized in that the statistical value is at least one of an average value, an intermediate value, a deviation, and a confidence interval width. The analysis system according to claim 1,
The storage unit further stores, as a random variable, the value of the indicator included in the health information of the individual who has performed the intervention, and stores a graphical model expressing a probabilistic dependency between the plurality of random variables,
The health transition prediction unit predicts the predicted transition with intervention using the graphical model and the health information of the candidate. An analysis method using an analysis system,
The analysis system includes a processor and a storage unit,
The storage unit indicates health information including at least one of an index indicating an individual's health level and an index indicating a factor affecting the health of the individual, and the individual who has performed an intervention supporting health. Store intervention information,
The analysis method is:
The processor acquires health information of a candidate who performs the intervention, and acquires time-series health information acquired before the intervention of the individual who has performed the intervention from the storage unit as a health transition with intervention. Based on the candidate's health information and the health transition with intervention, predicting the candidate's time-series health information when the candidate starts the intervention as a prediction transition with intervention, and A health transition prediction procedure for storing the predicted transition with intervention in the storage unit;
A duration prediction procedure in which the processor predicts a duration indicating a period in which the candidate's health is improved as an effect of the intervention using the predicted transition with intervention, and stores the duration in the storage unit When,
The analysis method characterized by including the subject selection procedure which selects the subject who performs the said intervention from the said candidate based on the said duration. The analysis method according to claim 9, comprising:
After the processor predicts a first duration of effect of the first intervention by the duration prediction procedure and selects a subject to perform the first intervention by the subject selection procedure; If the plan period during which an intervention is performed is longer than the first duration, the first duration has elapsed since the subject started the first intervention. Including a long-term planning procedure for predicting the health information of the subject based on the predicted transition with intervention,
The health transition prediction procedure is based on the predicted health information of the subject and the health transition with intervention after the subject has started a second intervention after the first duration has elapsed. Predicting time-series health information of the subject, and storing the predicted time-series health information in the storage unit,
The duration prediction procedure is configured to predict the presence of the intervention as a second duration of the effect of the second intervention, a period in which the candidate's health improves when the candidate starts the second intervention. Including a procedure to predict using transitions,
The long-term planning procedure performs the first intervention and the second intervention on the subject when the result of subtracting the first duration from the planning duration is equal to or less than the second duration. An analysis method characterized by including a procedure for determining what to do. The analysis method according to claim 9, comprising:
The health transition prediction procedure includes:
A procedure for acquiring time-series health information of individuals who have not performed the intervention from the storage unit as health transition without intervention;
Based on the candidate's health information and the health transition without intervention, a procedure for predicting the candidate's time-series health information when the candidate does not perform the intervention as a prediction transition without intervention;
Storing the prediction transition without intervention in the storage unit, and the duration prediction procedure includes a step of predicting the duration using the prediction transition with intervention and the prediction transition without intervention. Characteristic analysis method. The analysis method according to claim 11, comprising:
The health transition prediction procedure includes:
A procedure for acquiring health information at a predetermined timing when an individual who has performed the intervention starts intervention from the health transition with intervention as health information with intervention;
A procedure for identifying a first individual similar to the candidate from the individuals who performed the intervention based on the similarity between the candidate's health information and the health information with intervention;
Predicting the predicted transition with intervention based on the health transition with intervention of the first individual;
From the health transition without intervention, to obtain the health information of the earliest timing as health information without intervention;
Identifying a second individual similar to the candidate from individuals not performing the intervention based on the similarity between the candidate's health information and the health information without intervention;
And a procedure for predicting the predicted transition without intervention based on the health transition without intervention of the second individual. The analysis method according to claim 12, comprising:
The storage unit further stores state definition information for determining the health state of the individual from the health information,
The health transition prediction procedure includes:
Based on the state definition information, the health state determined from the health information with the intervention, the health state determined from the health information without the intervention, and the health state determined from the health information of the candidate are specified. Procedure and
A procedure for identifying the first individual based on the similarity between the health condition determined from the candidate's health information and the health condition determined from the identified health information with intervention;
And a procedure for identifying the second individual based on a similarity between a health condition determined from the candidate's health information and a health condition determined from the specified health information without intervention. Analysis method to do. The analysis method according to claim 12, comprising:
The health information includes a plurality of the indicators,
The analysis method includes a health index narrowing-down procedure for identifying an index of the health information that affects the effect of the intervention based on the health transition information with the intervention,
The health transition prediction procedure includes:
In the identified indicator, a procedure for identifying the first individual based on the similarity between the candidate's health information and the health information with intervention;
And a procedure for identifying the second individual based on a similarity between the candidate's health information and the health information without intervention in the identified index. The analysis method according to claim 9, comprising:
The health transition prediction procedure includes:
A procedure for acquiring time series health information at the start and after the start of the intervention of a plurality of individuals who have performed the intervention from the storage unit as health transition with intervention of the plurality of individuals,
A step of predicting the predicted transition with intervention by calculating a statistical value of health information for each of a plurality of timings included in the health transition with intervention of the plurality of individuals,
The analysis method is characterized in that the statistical value is at least one of an average value, an intermediate value, a deviation, and a confidence interval width.

Images