WO2023047766A1 - Risk analysis assistance system and risk analysis assistance method - Google Patents

Abstract

This risk analysis assistance system comprises: a processor; and a storage device connected to the processor, wherein the storage device stores health information about the health of a plurality of persons, attribute information on the plurality of persons, and a plurality of pieces of health state definition information, and the processor constructs a risk model for calculating the risk of changes in health states on the basis of the health information, the attribute information, and the plurality of pieces of health state definition information, calculates risk values indicating the risk of changes in the health states of the plurality of persons on the basis of the health information, the attribute information, and the risk model, and calculates, on the basis of the risk values, the order of priority for insurance guidance for the plurality of persons.

Description

Risk analysis support system and risk analysis support method Import by reference

This application claims the priority of Japanese Patent Application No. 2021-156885 filed on September 27, 2021, and incorporates the contents thereof into the present application by reference.

The present invention relates to technology that supports analysis of the risk of changes in health conditions.

Health insurers, which are operated based on health insurance premiums, are striving to reduce medical expenses and improve future income and expenditure by implementing health business measures that contribute to the maintenance and improvement of the health of subscribers. In order to maintain and improve the health of subscribers, it is necessary to secure human resources for those who provide health guidance, and it is difficult to provide services to all subscribers. Therefore, in order to obtain the maximum cost-effectiveness with limited resources, it is necessary to extract and select persons to be instructed. Until now, techniques for extracting subjects using information on medical expenses have been disclosed.

For example, Japanese Unexamined Patent Application Publication No. 2012-128670 (Patent Document 1) describes "a health business support system for selecting a person subject to health guidance based on medical insurance claim information, medical examination information, and health guidance information, including health insurance A medical cost model creation unit that creates a medical cost model that shows the predicted medical cost for each subscriber's severity and test value, and a test value improvement model that creates a test value improvement model that shows the amount of improvement for each severity and test value The creation unit, the predicted medical cost reduction effect calculation unit that calculates the predicted medical cost reduction amount due to health guidance for each severity and test value, and the health insurance subscribers belonging to the severity and test value with high predicted medical cost reduction amount and a target person selection unit for selecting a target person for health guidance.”

Patent document 1: JP 2012-128670

The method described in Patent Document 1 uses a method of estimating medical costs expected in the future to calculate the medical costs expected to be suppressed when test values improve due to health guidance. It extracts the target person according to the value. However, in this method, the extraction criterion is medical expenses, and measures against onset, hospitalization, and other event occurrence risks other than medical expenses are not considered. Therefore, in order to solve the above problems, the present invention makes it possible to perform effective target extraction by prioritizing guidance targets in consideration of risks related to multiple diseases or multiple severity levels of diseases. The purpose is to

In order to solve at least one of the above problems, a representative example of the invention disclosed in the present application is a risk analysis support system, comprising a processor and a storage device connected to the processor, The storage device holds health information related to the health of a plurality of persons, attribute information of the plurality of persons, and definition information of a plurality of health conditions, and the processor stores the health information, the attribute information and the constructing a risk model for calculating a risk of a change in the health condition based on the definition information of a plurality of health conditions; and based on the health information, the attribute information and the risk model, the plurality of persons It is characterized by calculating a risk value indicating the risk of a change in health condition, and calculating the priority of insurance guidance for the plurality of persons based on the risk value.

According to one aspect of the present invention, it is possible to appropriately determine the priority of health guidance for preventing changes in health conditions such as the onset or aggravation of diseases. Problems, configurations, and effects other than those described above will be clarified by the following description of the embodiments.

It is a block diagram showing an example of composition of a risk analysis support system of Example 1 of the present invention. FIG. 4 is an explanatory diagram showing an example of basic medical examination information managed by a medical examination history information management unit according to the first embodiment of this invention; It is explanatory drawing which shows an example of the disease name information which the medical examination history information management part of Example 1 of this invention manages. FIG. 4 is an explanatory diagram showing an example of medical practice information managed by a medical examination history information management unit according to Example 1 of the present invention; FIG. 4 is an explanatory diagram showing an example of health checkup information managed by a health checkup information management unit according to the first embodiment of this invention; FIG. 4 is an explanatory diagram showing an example of attribute information managed by an attribute information management unit according to Embodiment 1 of this invention; FIG. 4 is an explanatory diagram showing an example of severity definition information managed by an analysis data management unit according to Example 1 of this invention; FIG. 5 is an explanatory diagram showing an example of severity determination result information managed by an analysis data management unit according to Example 1 of this invention; FIG. 4 is an explanatory diagram showing an example of guidance history information managed by an analysis data management unit according to Example 1 of this invention; FIG. 4 is an explanatory diagram showing an example of risk model parameter information managed by a risk model information management unit according to Example 1 of this invention; FIG. 4 is an explanatory diagram showing an example of target person extraction information managed by a target person extraction information management unit according to the first embodiment of the present invention; FIG. 4 is a flowchart showing an example of processing executed by a state determination unit according to Example 1 of the present invention; FIG. It is a flowchart which shows an example of the process which the risk-model construction part of Example 1 of this invention performs. FIG. 4 is a flowchart showing an example of processing executed by a risk value calculation unit according to Example 1 of the present invention; FIG. It is a flowchart which shows an example of the process which the subject extraction part of Example 1 of this invention performs. FIG. 4 is an explanatory diagram showing an example of a user interface corresponding to processing of a state determination unit and a risk model construction unit according to Example 1 of the present invention; FIG. 4 is an explanatory diagram showing an example of a user interface corresponding to the processes of the risk value calculation unit and subject extraction unit according to the first embodiment of the present invention; It is a block diagram which shows an example of a structure of the risk-analysis support system of Example 2 of this invention. FIG. 10 is an explanatory diagram showing an example of target disease definition information managed by an analysis data management unit according to the second embodiment of this invention; FIG. 10 is an explanatory diagram showing an example of severity determination result information managed by an analysis data management unit according to Example 2 of the present invention; FIG. 10 is an explanatory diagram showing an example of model identification threshold information managed by a risk model information management unit according to the second embodiment of this invention; FIG. 10 is an explanatory diagram showing an example of risk correction result information managed by a risk model information management unit according to the second embodiment of this invention; FIG. 10 is an explanatory diagram showing an example of target person extraction information managed by a target person extraction information management unit according to the second embodiment of the present invention; It is a flowchart which shows an example of the process which the risk-model construction part of Example 2 of this invention performs. FIG. 11 is a flow chart showing an example of processing executed by a target person extraction unit 114 and a risk value correction unit 115 according to Example 2 of the present invention; FIG. FIG. 11 is an explanatory diagram showing an example of a user interface corresponding to processing of the state determination unit 112 and the risk model construction unit 111 according to Example 2 of the present invention; FIG. 10 is an explanatory diagram showing an example of a user interface corresponding to the processes of the risk value calculation unit, the subject extraction unit, and the risk value correction unit according to the second embodiment of the present invention;

Hereinafter, embodiments of the present invention will be described based on the drawings.

FIG. 1 is a block diagram showing an example of the configuration of the risk analysis support system 101 of Example 1 of the present invention.

The risk analysis support system 101 is a computer system, and includes an input unit 102 such as a keyboard and a mouse, an output unit 103 representing a display for outputting display data, a CPU (Central Processing Unit) 104, a memory 105, a communication unit 108, and a storage unit. A medium 106 is provided.

The risk analysis support system 101 has a risk model construction unit 111, a state determination unit 112, a risk value calculation unit 113, and a subject extraction unit 114. The functions of the risk model construction unit 111 to the subject extraction unit 114 are implemented by the CPU 104 executing a program stored in the storage medium 106 . When these programs are executed by CPU 104, at least a portion of them may be copied to memory 105 as needed.

A database 107 is connected to the risk analysis support system 101 . The database 107 has a medical examination history information management section 120 , a physical examination information management section 121 , an attribute information management section 122 , an analysis data management section 123 , a risk model information management section 124 and a subject extraction information management section 125 . In this embodiment, the medical examination history information management unit 120 manages medical examination basic information 200 (FIG. 2), injury or disease name information 300 (FIG. 3), and medical practice information 400 (FIG. 4), which will be described later. The health checkup information management unit 121 manages health checkup information 500 (FIG. 5), which will be described later. The attribute information management unit 122 manages attribute information 600 (FIG. 6), which will be described later. The analysis data management unit 123 manages severity definition information 700 (FIG. 7), severity determination result information 800 (FIG. 8), and guidance history information 900 (FIG. 9), which will be described later. The risk model information management unit 124 manages risk model parameter information 1000 (FIG. 10), which will be described later. The target person extraction information management unit 125 manages target person extraction information (FIG. 11), which will be described later.

Database 107 may be stored, for example, in a storage system connected to risk analysis support system 101 via a network, or may be stored in risk analysis support system 101 (for example, by being stored in storage medium 106). good too. When the database 107 is stored in a system outside the risk analysis support system 101, at least part of its contents may be copied to the storage medium 106 or memory 105 as required. Also, the entire system including the computer having the input unit 102, the output unit 103, the CPU 104, the memory 105 and the storage medium 106, and the database 107 may be called a risk analysis support system.

Also, the risk analysis support system 101 may be implemented by one computer having the configuration shown in FIG. 1, for example, or may be implemented by a plurality of computers. For example, the information held by the database 107 described above may be distributed and stored in a plurality of storage media 106 or memories 105, and the functions of the risk analysis support system 101 described above may be performed by a plurality of CPUs 104 of a plurality of computers. It may be executed in a distributed manner.

FIG. 2 is an explanatory diagram showing an example of basic medical examination information 200 managed by the medical examination history information management unit 120 according to the first embodiment of the present invention.

The basic consultation information 200 is information on the history of each person's consultation at a medical institution. This information may be collected from, for example, a medical bill prepared by a medical institution, but is not limited to this, and can be used as long as it indicates when and who received medical treatment at a medical institution.

The basic consultation information 200 includes an individual ID 201 that identifies each person, a medical examination history ID 202 that identifies the medical treatment that each person received in the past, a medical institution code 203 that indicates the medical institution that performed the medical treatment, and the year and month that the medical treatment was performed. medical treatment date 204 indicating the medical treatment, total points 205 indicating information on medical expenses corresponding to the medical treatment, consultation type 206 indicating information on the type of medical treatment received (for example, inpatient or outpatient), and the number of days required for medical treatment Includes number of treatment days 207 shown. In the case where the information on the medical examination history as described above is collected from the medical bill, the medical examination history ID 202 may identify the medical bill. With these pieces of information, it is possible to perform aggregation and analysis based on information indicating medical examination histories such as receipts.

FIG. 3 is an explanatory diagram showing an example of the disease name information 300 managed by the medical history information management unit 120 according to the first embodiment of the present invention.

The injury or disease name information 300 is information about an injury or disease extracted from the medical examination history information. A code 303, a main disease flag 304 given to a disease that has invested the most medical resources among multiple diseases, and an inspection to confirm whether the disease name is afflicted, and it is unconfirmed It includes a suspicion flag 305 that indicates status and the like.

In the example of FIG. 3, the value "1" of the primary disease flag 304 indicates that the disease is a primary disease. Also, the value "1" of the suspicion flag 305 indicates that the injury or disease is suspected.

For example, when one patient receives medical treatment for multiple diseases in one month, multiple sets of the disease name 302, the disease code 303, the main disease flag 304, and the suspicion flag 305 are associated with the same medical history ID 202. . Also, each record of the injury or disease name information 300 is associated with the individual ID 201 via the medical history ID 202 . By using the disease name information 300, it becomes possible to analyze according to the disease.

FIG. 4 is an explanatory diagram showing an example of the medical practice information 400 managed by the medical examination history information management unit 120 according to the first embodiment of the present invention.

The medical practice information 400 is information on the medical practice performed on the patient in each month, extracted from the information on the medical examination history. From the medical practice name 402 indicating the medical practice performed, the medical practice code 403 corresponding to the medical practice, the medical practice score 404 determined for each medical practice, and the day information 405 showing the day on which the medical practice was performed It contains information 408 for the 31st. Each record of the medical practice information 400 is associated with an individual ID 201 via a medical examination history 202 .

Information 405 to 31st day information 408 indicating the date on which the medical practice was performed indicates whether or not the medical practice indicated by the medical practice name 402 was performed on each day from the 1st to the 31st of the month. is information indicating FIG. 4 shows 1st day information 405, 2nd day information 406, 3rd day information 407, and 31st day information 408 as an example, but actually information from 4th day to 30th day is also included. This information allows analysis by intervention.

FIG. 5 is an explanatory diagram showing an example of health checkup information 500 managed by the health checkup information management unit 121 according to the first embodiment of the present invention.

The health checkup information 500 is information about the result of a health checkup (medical checkup) that each person received, and includes a personal ID 201, a health checkup ID 502 that specifies the health checkup, a health checkup year 503 that indicates the year in which the health checkup was performed, and a BMI ( Body Mass Index) 504, fasting blood sugar 505, HbA1c 506, creatinine 507, interview result 508, and the like. The interview result 508 may include, for example, information indicating whether the person has a habit of drinking alcohol, whether or not he has a habit of exercising, and the like. The above-mentioned BMI 504 to creatinine 507 are typical examples of information obtained as a result of a health checkup, and in practice the health checkup information 500 may not include at least one of these items, or may include items other than these. Information (eg, systolic blood pressure, diastolic blood pressure, etc.) may also be included. This information allows analysis based on health status.

FIG. 6 is an explanatory diagram showing an example of attribute information 600 managed by the attribute information management unit 122 according to the first embodiment of this invention.

The attribute information 600 is information about the attributes of each person, and includes a personal ID 201, a gender 602, a date of birth 603, a subscription date 604 indicating information on the date of subscription to the health insurance of each person, and a date of subscription from the health insurance. It includes a date of withdrawal 605 indicating information on the date of withdrawal. This information enables analysis according to each person's age, gender, years of follow-up, and the like.

FIG. 7 is an explanatory diagram showing an example of severity definition information 700 managed by the analysis data management unit 123 according to the first embodiment of this invention.

The severity definition information 700 is definition information for judging the severity based on each person's medical examination history information, health checkup information, etc., and includes a definition ID 701, definition content 702, and judgment level 703. When information obtained from a person's medical examination history, medical examination results, etc. corresponds to the value of the definition content 702, the value of the determination level 703 is determined as the severity of the person. In this embodiment, four severity levels are defined from "0" indicating no onset to "3" indicating the most severe symptoms.

For example, the first line of the severity definition information 700 shown in FIG. 7 indicates that the severity level is determined to be "0" when the fasting blood sugar is less than 126 and the HbA1c is less than 6.5, The second row indicates that the severity level is determined to be "1" when the fasting blood sugar is 126 or higher and HbA1c is 6.5 or higher. The third line indicates that the severity level is determined to be "2" when diabetic complications are present. Line 4 indicates that the severity level is determined to be "3" for chronic renal failure. For example, the fasting blood sugar and HbA1c of each person may be obtained from the health checkup information 500, and whether or not each person has diabetes complications and whether or not each person has chronic renal failure may be obtained from basic consultation information. and may be obtained from the injury name information. The severity definition information 700 makes it possible to determine the severity of each person based on medical history information, health checkup information, and the like.

FIG. 8 is an explanatory diagram showing an example of severity determination result information 800 managed by the analysis data management unit 123 according to the first embodiment of this invention.

The severity determination result information 800 is information indicating the result of determining the severity based on each person's medical examination history information, medical examination information, etc., and the severity definition information 700. It includes severity level 0 matching status 803 , severity level 1 matching status 804 , severity level 2 matching status 805 , severity level 3 matching status 806 and determination result 807 .

The personal ID 201 is information that identifies a person. The target year 802 indicates a target year for severity determination. Severity level 0 applicable/not applicable 803 to severity level 3 applicable/not applicable 806 indicate that information such as medical examination history and medical checkup result of each person in the target year is defined by the severity level definition information 700. It indicates whether or not it corresponds to 3. In the example of FIG. 8, "1" indicates applicable and "0" indicates non-applicable. The determination result 807 indicates the severity level determined based on the information of severity level 0 matching 803 to severity level 3 matching 806 .

For example, the first line of the severity determination result information 800 shown in FIG. , and as a result, the severity level of the person in question for the year in question was determined to be "0." Also, the 10th to 12th lines show the judgment results of the person with the personal ID “P003” for the fiscal years 2019 to 2021. In 2019, the medical history and medical examination results correspond to severity level 1, not to severity levels 2 and 3, and as a result, the highest "1" among the applicable severity levels is the person's It is determined as the severity level of the year. In 2020, the medical examination history and medical examination results correspond to severity levels 1 and 2, but not to severity level 3, and as a result, the highest "2" among the applicable severity levels is the person's It is determined as the severity level of the year. In 2021, the medical examination history and medical examination results fall under severity levels 1 to 3, and as a result, "3", which is the highest among the applicable severity levels, is determined as the severity level of the person in question for that year. there is

FIG. 9 is an explanatory diagram showing an example of the guidance history information 900 managed by the analysis data management unit 123 according to the first embodiment of this invention.

Guidance history information 900 is information for managing the history of health guidance given to each person in the past, and includes an individual ID 201, target year 902, whether or not severity level 1 guidance has been provided 903, whether or not severity level 2 guidance has been provided. 904 , severity level 3 guidance implementation presence/absence 905 , guidance non-target flag 906 and guidance target determination reason 907 .

The personal ID 201 is information that identifies a person. The target year 902 indicates the target year of the health guidance. Severity level 1 guidance implementation/non-execution 903 to severity level 3 guidance implementation/non-execution 905 indicate whether health guidance corresponding to each severity level has been provided to each person. Guidance exclusion flag 906 indicates whether each person is excluded from insurance guidance. Guidance target determination reason 907 indicates the reason when it is determined that each person is not subject to insurance guidance.

For example, the first line of the guidance history information 900 shown in FIG. 9 indicates that the person with the personal ID "P001" received no insurance guidance corresponding to any severity level in FY2020. In addition, the fourth line shows that the person with the personal ID “P002” received insurance guidance corresponding to severity level 1 in fiscal 2019, and that the same insurance guidance was given to the person for multiple years. Indicates that the person is currently not subject to health guidance because it was performed. In addition, the fifth line indicates that the person with the personal ID “P003” received insurance guidance corresponding to severity levels 2 and 3 in 2021, and that the person has already been treated by a medical institution. Indicates that the person is currently not subject to health guidance because it has been started.

If the purpose of insurance guidance is to control medical expenses by preventing the onset and aggravation of illness, rather than giving the same guidance to the same person over and over again, it is better to give guidance to people who have never given guidance before. A higher effect can be expected by doing In addition, for those who have already been severely ill and have started treatment at a medical institution, the goal of curtailing medical expenses has already become unfulfilled. Therefore, by excluding such cases from insurance guidance, it is possible to contribute to curtailing medical expenses using limited resources.

FIG. 10 is an explanatory diagram showing an example of risk model parameter information 1000 managed by the risk model information management unit 124 according to the first embodiment of this invention.

The risk model parameter information 1000 includes a model ID 1001 that identifies a model for assessing risk (risk model), a model name 1002 that indicates what kind of model each model is, and model parameters that indicate the structure and parameters of the model. 1003 included.

For example, the risk model construction unit 111, based on the basic consultation information 200, the disease name information 300, the medical treatment information 400, the medical examination information 500, the attribute information 600, the severity definition information 700 and the severity determination result information 800, each One or more models are generated for calculating the risk value of change in the health condition of each person from the value of at least one of the person's medical examination history, health checkup results, and attributes. The type, structure and parameters of the generated risk model are stored in risk model parameter information 1000 .

Here, the risk value of a change in health condition may be, for example, a value indicating the risk of developing any disease, or a value indicating the risk of a change in the severity of any disease (especially aggravation). There may be. That is, as the risk model parameter information 1000, for example, a risk model for calculating the onset risk for each type of disease may be held, or a risk model for calculating the risk of change in severity of a specific disease. may be retained. In this example, a risk model is maintained for calculating the risk of developing and changing severity of a particular disease (eg diabetes). Hereinafter, this embodiment will be described by taking as an example the case of calculating the risk of the onset of a specific disease and the change in severity. Needless to say.

In the example of FIG. 10, as the risk model with the model ID "1", a model for calculating the risk that a person who has not yet developed a specific disease (for example, diabetes) will develop the disease with a severity level of 1 or higher. is retained. In this example, parameters of a model having age, fasting blood sugar, etc. as explanatory variables and probability of developing a disease with a severity level of 1 or higher as an objective variable are registered.

Similarly, as the risk model with model ID "2", a model for calculating the risk of a person with severity level 1 or lower developing the disease with severity level 2 or higher is held. In this example, parameters of a model having age, sex, fasting blood sugar, creatinine, etc. as explanatory variables and probability of developing a disease with a severity level of 2 or higher as an objective variable are registered. Furthermore, a model for calculating the risk that a person with a severity level of 2 or less develops the disease with a severity level of 3 or higher is held as a risk model with a model ID of "3". In this example, parameters of a model having age, sex, urinary protein, creatinine, etc. as explanatory variables and the probability of developing a disease with a severity level of 3 or higher as an objective variable are registered.

Although this example shows a model for calculating the probability of occurrence of a severity level above a certain level, a model for calculating the probability of occurrence of a severity level 1 disease, for example, may A model may be created to calculate the probability of developing a disease. As model parameters, for example, parameters of a regression model may be registered, but parameters of other models may be registered as long as the model can predict risk values. This information makes it possible to calculate the risk of developing the disease for each individual severity level.

FIG. 11 is an explanatory diagram showing an example of the target person extraction information 1100 managed by the target person extraction information management unit 125 according to the first embodiment of the present invention.

The subject extraction information 1100 is information for managing each person's onset risk calculated based on the risk model parameter information 1000 and the intervention priority calculated based thereon. FIG. 11 shows target person extraction information 1100 consisting of tables 1110, 1120 and 1130 as an example.

A table 1110 holds the probability of developing a disease with a severity level of 3 or higher for a person whose current severity level is "2", and the order of priority for intervention such as health guidance calculated based on this probability. Specifically, the individual ID 1111 is information that identifies each person. Current level 1112 indicates each person's current severity level (level "2" in this example). The level 3 or higher onset probability 1113 indicates, for example, the onset probability of severity level 3 or higher for each person calculated using the risk model with the model ID “3” of the risk model parameter information 1000 . The intervention priority 1114 indicates the priority of intervention such as insurance guidance for each person. In this example, a high intervention priority is given to a person with a high onset probability.

A table 1120 holds the probability of developing a disease with a severity level of 2 or higher for a person whose current severity level is "1", and the intervention priority order calculated based thereon. Specifically, the personal ID 1121 is information that identifies each person. Current level 1122 indicates each person's current severity level (level "1" in this example). The level 2 or higher onset probability 1123 indicates, for example, the onset probability of severity level 2 or higher for each person calculated using the risk model with the model ID “2” of the risk model parameter information 1000 . The intervention priority 1124 indicates the priority of intervention such as insurance guidance for each person. In this example, a high intervention priority is given to a person with a high onset probability.

A table 1130 holds the probability of developing a disease with a severity level of 1 or higher for a person whose current severity level is "0" (that is, has not yet developed symptoms), and the intervention priority order calculated based thereon. Specifically, the personal ID 1131 is information that identifies each person. Current level 1132 indicates each person's current severity level (level "0" in this example). The level 1 or higher onset probability 1133 indicates, for example, the onset probability of severity level 1 or higher for each person calculated using the risk model with the model ID “1” of the risk model parameter information 1000 . The intervention priority 1134 indicates the priority of intervention such as insurance guidance for each person. In this example, a high intervention priority is given to a person with a high onset probability.

In the example of FIG. 11, since the intervention priority is set for each table, there will be multiple persons with the highest intervention priority, for example. In this case, which severity level of human intervention is ultimately prioritized depends on, for example, the type and amount of medical resources for treatment and the type and amount of resources for health guidance. can decide. For example, a person with a high intervention priority based on the probability of developing severity level 3 may be given the highest priority.

Also, in the example of FIG. 11, the target person extraction information 1100 is shown divided into three tables for the sake of explanation, but it may actually be one table. For example, the individual ID of each person, the current severity level of each person, the probability of occurrence of the severity level one step higher than the current severity level of each person, and the intervention of each person calculated based on it A single table containing priority and may be held as the target person extraction information 1100 .

FIG. 12 is a flow chart showing an example of processing executed by the state determination unit 112 according to the first embodiment of the present invention.

When the process is started (step 1201), the state determination unit 112 reads attribute information, health checkup information, medical history information, and severity definition (steps 1202 to 1205). As a result, for example, attribute information 600, health checkup information 500, basic consultation information 200, disease name information 300, medical practice information 400, and severity definition information 700 are read.

Next, the state determination unit 112 generates a severity flag for each year for each person based on the read information (step 1206). As a result, for example, the values of the severity level 0 matching presence/absence 803 to the severity level 3 matching presence/absence 806 of the severity determination result information 800 are generated.

Next, the state determination unit 112 generates severity determination result information for each year for each person based on the severity flag generated in step 1206 (step 1207). As a result, for example, the value of the determination result 807 of the severity determination result information 800 is generated, and the severity determination result information 800 is completed.

This completes the processing (step 1208).

FIG. 13 is a flow chart showing an example of processing executed by the risk model construction unit 111 according to the first embodiment of the present invention.

When the process starts (step 1301), the risk model building section 111 sets the conditions for building the risk model (step 1302). As a result, for example, the temporal range of explanatory variables and objective variables, the age range of target persons, and the like are set.

Next, the risk model construction unit 111 reads attribute information, health checkup information, medical examination history information, and severity determination results (steps 1303 to 1306). As a result, for example, attribute information 600, health checkup information 500, basic consultation information 200, disease name information 300, medical practice information 400, and severity determination result information 800 are read.

Next, the risk model construction unit 111 creates an analysis data set based on the information read in steps 1303-1306 and the conditions set in step 1302 (step 1307). For example, the risk model construction unit 111 extracts information that meets the conditions set in step 1302 from the information read in steps 1303 to 1306, and compares it based on the person's ID to obtain the objective variable. An analysis data set is created that includes values (in this example, severity determination results) and their corresponding explanatory variable values.

Next, the risk model construction unit 111 extracts data for model construction based on the analysis data set created in step 1307 (step 1308), and constructs a risk model (step 1309). Parameters of the constructed model are held as risk model parameter information 1000 .

This completes the processing (step 1310).

FIG. 14 is a flow chart showing an example of processing executed by the risk value calculation unit 113 according to the first embodiment of the present invention.

When the process starts (step 1401), the risk value calculation unit 113 reads attribute information, health checkup information, and medical examination history information (steps 1402-1404). As a result, for example, attribute information 600, health checkup information 500, basic consultation information 200, disease name information 300, and medical practice information 400 are read.

Next, the risk value calculation unit 113 constructs an evaluation data set based on the read information (step 1405). For example, from the information read in steps 1402 to 1404, the risk value calculation unit 113 extracts data corresponding to explanatory variables of each person whose risk value is to be evaluated and intervention priority is to be determined, Build the data to feed into the risk model.

Next, the risk value calculation unit 113 determines the current severity level of each person based on the constructed evaluation data set, and determines the risk model to be applied to each person (step 1406).

Next, the risk value calculation unit 113 applies the risk model determined in step 1406 to the evaluation data set constructed in step 1405 to obtain a risk value for each severity of each target person (that is, the severity level). onset probability) is calculated (step 1407).

Next, the risk value calculator 113 stores the risk value calculated in step 1407 (step 1408). For example, the calculated risk value is stored in any of the level 3 or higher onset probability 1113 , level 2 or higher onset probability 1123 , or level 1 or higher onset probability 1133 of the subject extraction information 1100 .

This completes the processing (step 1409).

FIG. 15 is a flow chart showing an example of processing executed by the subject extraction unit 114 according to the first embodiment of the present invention.

When the process starts (step 1501), the subject extraction unit 114 sets extraction conditions (step 1502). For example, the age range of the target person may be set, the number of people with the highest priority for intervention to be extracted, the number of people to be extracted for each severity level, and the number of people with high severity levels to be extracted. A condition may be set such as which of the low-ranked persons is given priority.

Next, the subject extraction unit 114 reads the risk evaluation result (step 1503). For example, the onset probability of each severity level for each person calculated by the risk value calculation unit 113 is read.

Next, the target person extraction unit 114 adds the priority order for intervention such as health guidance to the risk assessment results read in step 1503 (step 1504). For example, when risk evaluation results such as level 3 or higher onset probability 1113, level 2 or higher onset probability 1123, and level 1 or higher onset probability 1133 of subject extraction information 1100 are read, intervention priorities 1114, 1124 and 1134 priority is added.

Next, the target person extraction unit 114 stores the priority added in step 1504 as the target person extraction information 1100 (step 1505).

This completes the processing (step 1506). As a result, for example, a person with a high risk value is given a high priority, and insurance guidance is given preferentially.

Note that the subject extraction unit 114 identifies the trend of change in risk values not only from the newly calculated risk values as described above, but also from previously calculated risk values and newly calculated risk values. The results may be reflected in prioritization decisions. For example, the subject extraction information management unit 125 manages the calculation results of the past risk values, and based on the past risk values and the newly calculated risk values, the risk value increase trend exceeds a predetermined condition. Persons may be prioritized for intervention. Thereby, for example, even if the current risk value is not so high, a high priority is added to a person whose increasing trend is remarkable.

Next, an example of a user interface provided by the risk analysis support system 101 when executing the processes shown in FIGS. 12 to 15 will be described with reference to FIGS. 16 and 17. FIG.

FIG. 16 is an explanatory diagram showing an example of a user interface corresponding to the processing of the state determination unit 112 and the risk model construction unit 111 of Example 1 of the present invention.

A risk model building screen 1600 shown in FIG. 16 is an example of display data output by the risk analysis support system 101. For example, a data reading unit 1601, a model building condition setting unit 1602, a model building process execution button 1603, analysis data A set creation result display section 1604 and a model construction result display section 1605 are included.

Note that this display data may be output as an image by the output unit 103 or may be output by the communication unit 108 . In the latter case, the display data is transferred from the communication unit 108 via a network (not shown) to an external device (for example, a terminal device used by a user, not shown), and output as an image by the external device. good. In that case, the information input via the screen is input using the input unit (not shown) of the external device and input to the risk analysis support system 101 via the network and communication unit 108 . The same applies to other screens to be described later.

In the data reading unit 1601, conditions for data to be read for state determination are specified. For example, it is possible to specify the definition of the condition for judging the health condition (specifically, the condition for judging the severity level, for example), the attribute information of the target person, the result of the medical checkup of the target, the history of the medical examination of the target, and the like. Reading in steps 1202 to 1205 is performed based on the conditions specified here.

In the model building condition setting unit 1602, conditions for building a risk model are set. For example, it is possible to specify the temporal range of data used as explanatory variables, the temporal range of data used as objective variables, and the age range of the target person. These conditions are designated in step 1302, and data that meets the designated conditions is read in steps 1303-1306.

When the model construction processing execution button 1603 is operated, the processing of the state determination unit 112 and the risk model construction unit 111 is executed according to the conditions specified by the data reading unit 1601 and model construction condition setting unit 1602 .

The analysis data set created in step 1307 is displayed in the analysis data set creation result display section 1604 . Specifically, for each person, for example, the level of severity as an explanatory variable, attributes such as gender and age, information such as BMI and interview results obtained from medical checkup results, etc., and presence or absence of past diseases obtained from medical history information , and the severity level, which is the objective variable, etc., are displayed.

The model construction result display section 1605 displays information on the model constructed by the risk model construction section 111 . For example, information corresponding to the risk model parameter information 1000 may be displayed.

FIG. 17 is an explanatory diagram showing an example of a user interface corresponding to the processing of the risk value calculation unit 113 and the subject extraction unit 114 according to the first embodiment of the present invention.

A target person extraction screen 1700 shown in FIG. 17 is an example of display data output by the risk analysis support system 101. For example, a data reading unit 1701, a target person extraction condition setting unit 1702, a target person extraction execution button 1703, and a target A person extraction result display portion 1704 is included.

The data reading unit 1701 is the same as the data reading unit 1601 in FIG.

In the subject extraction condition setting unit 1702, conditions for extracting intervention subjects are set. For example, the temporal range of data used as explanatory variables and the age range of the person of interest can be specified.

When the target person extraction execution button 1703 is operated, the processes of the risk value calculation unit 113 and the target person extraction unit 114 are executed according to the conditions specified by the target person extraction condition setting unit 1702 .

The target person extraction result display unit 1704 displays the results of the processing of the risk value calculation unit 113 and the target person extraction unit 114 . For example, information corresponding to the target person extraction information 1100 may be displayed.

According to Example 1 above, it is possible to appropriately determine the priority of health guidance for preventing changes in health conditions such as the onset or aggravation of diseases. For example, by appropriately selecting people at high risk of onset or at high risk of aggravation and providing preferential insurance guidance, it is expected that limited resources will be utilized to curb an increase in medical costs.

Next, Example 2 of the present invention will be described. Except for the differences described below, the parts of the system of Example 2 have the same functions as the like-numbered parts of Example 1 shown in FIGS. are omitted.

FIG. 18 is a block diagram showing an example of the configuration of the risk analysis support system 101 of Example 2 of the present invention.

In the risk analysis support system 101 of the second embodiment, the storage medium 106 further has a risk value correction unit 115. In addition, there are differences in the processing of the risk model construction unit 111 and the target person extraction unit 114, which will be described later. In the database 107, the information managed by the analysis data management unit 123, the risk model information management unit 124, and the subject extraction information management unit 125 has differences, which will be described later. Note that the function of the risk value correction unit 115 is realized by executing the program stored in the storage medium 106 by the CPU 104, like the functions of the other units.

In addition, in Example 1, the onset risk for each person's severity level was calculated for a specific disease, and the intervention priority was calculated based on this. On the other hand, in Example 2, the onset risk of each disease for each person is calculated for a plurality of diseases, and the intervention priority is calculated based on this. Both the onset risk for each severity level of a specific disease in Example 1 and the onset risk for each disease in Example 2 are examples of the risk of changes in a person's health condition. It goes without saying that Example 2 below can also be applied to the risk of developing a particular disease for each severity level.

FIG. 19 is an explanatory diagram showing an example of target disease definition information 1900 managed by the analysis data management unit 123 according to the second embodiment of the present invention.

The target disease definition information 1900 is definition information for determining whether or not each person has developed a disease based on each person's medical examination history information, health checkup information, etc., and definition ID 1901 that identifies each definition. , a target disease name 1902 specifying the target disease of the definition and an ICD10 definition 1903 indicating the contents of the definition. The ICD10 definition 1903 describes the code of ICD10 (International Classification of Diseases, 10th edition). For example, a code indicating information defining type 2 diabetes, cardiovascular disease, cerebrovascular disease, etc. is described as the ICD10 definition 1903 . In the second embodiment, the above three diseases are described for the sake of explanation, but in reality the target disease definition information 1900 can include more disease definition information, and many diseases can be detected in the processing described later. The presence or absence of onset can be determined. The target disease definition information 1900 makes it possible to determine whether or not each person has developed each disease based on medical examination history information, medical examination information, and the like.

FIG. 20 is an explanatory diagram showing an example of severity determination result information 2000 managed by the analysis data management unit 123 according to the second embodiment of the present invention.

The severity determination result information 2000 is information indicating the result of determining whether or not each person has developed each disease based on each person's medical history information, medical examination information, etc., and the target disease definition information 1900. It includes an ID 201, a target year 2002, type 2 diabetes mellitus 2003, cardiovascular disease 2004, and cerebrovascular disease 2005.

The personal ID 201 is information that identifies a person. The target year 2002 indicates the target year of determination. Type 2 diabetes mellitus 2003, cardiovascular disease 2004, and cerebrovascular disease 2005 are defined by the target disease definition information 1900 for each person's medical examination history and medical checkup results in the target year. Indicates whether type 2 diabetes, cardiovascular disease, and cerebrovascular disease are applicable. In the example of FIG. 20, "1" indicates applicable and "0" indicates non-applicable.

FIG. 21 is an explanatory diagram showing an example of model identification threshold information 2100 managed by the risk model information management unit 124 according to the second embodiment of the present invention.

The model identification threshold information 2100 is calculated using a model ID 2101 that identifies a risk model, a model name 2102 that indicates what kind of model each model is, model parameters 2103 that indicate the structure and parameters of the model, and the risk model. A discrimination threshold value 2104 is included for determining the presence or absence of onset of each disease based on the calculated risk value.

For example, the risk model construction unit 111, based on the basic consultation information 200, the disease name information 300, the medical treatment information 400, the medical examination information 500, the attribute information 600, the target disease definition information 1900 and the severity determination result information 2000, each Generate one or more models for calculating the risk value of changes in the health condition of each person (onset of each disease in Example 2) from the value of at least one of the person's medical examination history, medical examination results, and attributes do. The type, structure, parameters, and the like of the generated risk model are stored in the model identification threshold information 2100 .

In the example of FIG. 21, a model for calculating the risk of developing type 2 diabetes is held as the risk model with model ID "1". In this example, parameters of a model with age, sex, fasting blood sugar, HbA1c, etc. as explanatory variables and the probability of developing type 2 diabetes as objective variables are registered, and 0.19 is registered as a discrimination threshold. This indicates that type 2 diabetes is determined to be present when the probability of occurrence calculated by the model exceeds 0.19.

Similarly, a model for calculating the risk of developing cardiovascular disease is held as a risk model with model ID "2". In this example, parameters of a model with age, gender, systolic blood pressure, diastolic blood pressure, etc. as explanatory variables and the probability of developing cardiovascular disease as objective variables are registered, and 0.20 is registered as a discrimination threshold. there is Furthermore, a model for calculating the risk of developing a cerebrovascular disease is held as a risk model with model ID "3". In this example, parameters of a model with age, gender, systolic blood pressure, triglycerides, etc. as explanatory variables and the probability of developing a cerebrovascular disease as an objective variable are registered, and 0.02 is registered as a discrimination threshold. there is

The method of setting the value of the discrimination threshold 2104 is not limited, but it is desirable to set it so that the discrimination performance is maximized. For example, the threshold may be set at a point where sensitivity + specificity - 1 is maximum on an ROC (Receiver Operating Characteristic) curve. The discriminative threshold information by model 2100 makes it possible to calculate the risk of developing each disease for each person.

FIG. 22 is an explanatory diagram showing an example of risk correction result information 2200 managed by the risk model information management unit 124 according to the second embodiment of the present invention.

The risk correction result information 2200 is information indicating the result of calculating the risk of developing each disease for each person based on the risk model and the result of correcting it based on the identification threshold. Specifically, the risk correction result information 2200 includes an individual ID 201, a type 2 diabetes risk value 2202, a cardiovascular disease risk value 2203, a cerebrovascular disease risk value 2204, a corrected type 2 diabetes risk value 2205, a cardiovascular A disease onset corrected risk value 2206, a cerebrovascular disease onset corrected risk value 2207, and a disease 2208 with the highest risk of onset are included.

The personal ID 201 is information that identifies a person. A type 2 diabetes risk value 2202, a cardiovascular disease risk value 2203, and a cerebrovascular disease risk value 2204 are values indicating the risk of each person developing type 2 diabetes calculated based on the corresponding risk model. (e.g. probability of onset), a value indicating the risk of developing cardiovascular disease (e.g. probability of onset), and a value indicating the risk of developing cerebrovascular disease (e.g. probability of onset).

Type 2 diabetes onset corrected risk value 2205, cardiovascular disease onset corrected risk value 2206 and cerebrovascular disease onset corrected risk value 2207 are, respectively, type 2 diabetes onset risk value 2202, cardiovascular disease onset risk value 2203 and cerebrovascular disease onset It is a value obtained by correcting the risk value 2204 based on the corresponding discrimination threshold. A disease with the highest onset risk 2208 indicates a disease determined to have the highest onset risk based on the corrected risk value.

For example, the first line of the risk correction result information 2200 shown in FIG. It shows that the risk value corrected based on the corresponding discrimination threshold of "0.19" is "0.06". In this example, the correction is performed by subtracting the discrimination threshold from the onset risk value calculated using the risk model.

Similarly, in the first line, the cardiovascular disease risk value “0.24” of the person with the personal ID “P001” is corrected to “0.04” based on the discrimination threshold “0.20”, It shows that the disease onset risk value “0.23” is corrected to “0.21” based on the discrimination threshold “0.02”. As a result, the disease with the highest onset risk for the person with the personal ID “P001” is determined to be cerebrovascular disease based on the corrected risk value.

In general, for diseases with a low frequency of occurrence, the probability of occurrence, which is the output of the risk model, is lower than for diseases that do not occur, so it is difficult to assess risk simply by comparing the probability of occurrence between diseases. On the other hand, if the onset probability is low, the identification threshold for determining whether or not the onset will occur is also low. It becomes possible to compare the onset risk between diseases.

For example, in the above example, the risk values for developing type 2 diabetes and cardiovascular disease for the person with the personal ID “P001” are both slightly higher than the identification threshold, but the risk value for developing cerebrovascular disease is higher than the identification threshold. significantly higher. Therefore, among these three diseases, the person has the highest risk of developing cerebrovascular disease (that is, there is a high need for intervention such as insurance guidance to prevent the onset of cerebrovascular disease). It can be said. However, when simply comparing the probability of developing type 2 diabetes, which is the output of the risk model, it is determined that the risk of developing type 2 diabetes is the highest, and insurance guidance or the like is provided to prevent the onset of type 2 diabetes.

On the other hand, in the present embodiment, it is possible to determine that the risk of developing cerebrovascular disease is the highest by using the value obtained by correcting the probability of occurrence using the discrimination threshold as described above as the risk value. become.

In the above example, correction is performed by subtracting the discrimination threshold from the onset risk value calculated using the risk model. At least one of a correction to lower the value and a correction to increase the disease onset risk value with a relatively low discrimination threshold may be performed.

FIG. 23 is an explanatory diagram showing an example of the target person extraction information 2300 managed by the target person extraction information management unit 125 according to the second embodiment of the present invention.

The subject extraction information 2300 is the onset risk of each person calculated based on the risk model parameter information 1000, the onset risk corrected based on the discrimination threshold, and the intervention calculated based on the corrected onset risk. This is information for managing the order of priority. FIG. 23 shows target person extraction information 2300 consisting of tables 2310, 2320 and 2330 as an example.

Table 2310 holds type 2 diabetes onset risk, onset risk corrected based on the discrimination threshold, and priority for intervention such as health guidance calculated based on the corrected onset risk. be done. Specifically, the personal ID 2311 is information that identifies each person. The target disease 2312 indicates a disease (type 2 diabetes in this example) for which the onset risk and intervention priority are calculated based on the risk. The onset risk 2313 indicates, for example, the probability of type 2 diabetes onset for each person calculated using the risk model with the model ID “1” in the model-specific identification threshold information 2011 . The corrected onset risk 2314 indicates the onset risk obtained by correcting the type 2 diabetes onset probability of each person using the value "0.19" of the discrimination threshold 2104 of the model. The intervention priority 2315 indicates the priority of intervention such as insurance guidance for each person. In this example, high intervention priority is given to persons with a high risk of developing the disease after correction.

Table 2320 holds the risk of developing cardiovascular disease, the risk of developing cardiovascular disease corrected based on the discrimination threshold, and the priority of intervention such as health guidance calculated based on the corrected risk of developing. be done. Specifically, the personal ID 2321 is information that identifies each person. The target disease 2322 indicates a disease (cardiovascular disease in this example) for which the onset risk and intervention priority are calculated based on the risk. The onset risk 2323 indicates, for example, the probability of onset of cardiovascular disease for each person calculated using the risk model with the model ID “2” in the model-specific identification threshold information 2011 . The corrected onset risk 2324 indicates the onset risk obtained by correcting the cardiovascular disease onset probability of each person using the value "0.20" of the discrimination threshold 2104 of the model. The intervention priority 2325 indicates the priority of intervention such as insurance guidance for each person. In this example, high intervention priority is given to persons with a high risk of developing the disease after correction.

Table 2330 holds the risk of developing cerebrovascular disease, the risk of developing it corrected based on the discrimination threshold, and the priority of intervention such as health guidance calculated based on the corrected risk of developing. be done. Specifically, the personal ID 2331 is information that identifies each person. The target disease 2332 indicates a disease (cerebrovascular disease in this example) for which the onset risk and intervention priority are calculated based on the risk. The onset risk 2333 indicates, for example, the probability of onset of cerebrovascular disease for each person calculated using the risk model with the model ID “3” of the model-specific identification threshold information 2011 . The corrected onset risk 2334 indicates the onset risk obtained by correcting the onset probability of each person's cerebrovascular disease using the value "0.02" of the discrimination threshold 2104 of the model. The intervention priority 2335 indicates the priority of intervention such as insurance guidance for each person. In this example, high intervention priority is given to persons with a high risk of developing the disease after correction.

It should be noted that the table 2310 may include only information about a person determined to have the highest risk of developing type 2 diabetes based on the corrected risk of developing type 2 diabetes. Similarly, table 2320 may include only information about persons determined to be at the highest risk of developing cardiovascular disease based on the corrected risk of incidence, and table 2330 may include Based on risk, only information about persons determined to be at highest risk of developing cerebrovascular disease may be included.

Next, the processing executed by the risk analysis support system 101 of the second embodiment will be described with reference to a flowchart. The processing executed by the state determination unit 112 of the second embodiment is similar to that of the first embodiment (see FIG. 12). However, the state determination unit 112 reads the target disease definition information 1900 in step 1205 , determines whether or not each disease is applicable in step 1206 , and holds the result as the severity determination result information 2000 .

FIG. 24 is a flow chart showing an example of processing executed by the risk model construction unit 111 according to the second embodiment of the present invention.

When the process starts (step 2401), the risk model building section 111 sets the conditions for building a risk model (step 2402), and reads attribute information, health checkup information, and medical examination history information (steps 2403-2405). These are the same as steps 1302 to 1305 (FIG. 13) of the first embodiment.

Next, the risk model construction unit 111 reads the severity determination result information 2000 (step 2406).

Next, the risk model construction unit 111 creates an analysis data set based on the information read in steps 2403-2406 and the conditions set in step 2402 (step 2407). For example, the risk model construction unit 111 extracts information that meets the conditions set in step 2402 from the information read in steps 2403 to 2406, and compares it based on the person's ID to obtain the objective variable. An analysis data set is created that includes values (in this example, the presence or absence of each disease) and the corresponding explanatory variable values.

Next, the risk model construction unit 111 extracts data for model construction based on the analysis data set created in step 2407 (step 2408), and constructs a risk model (step 2409). Parameters of the constructed model are held as model parameters 2103 of the model identification threshold information 2100 .

Next, the risk model construction unit 111 calculates a discrimination threshold (step 2410). As described above, the method for calculating this threshold is not limited, but it is desirable to set it so as to maximize the discrimination performance. The calculated identification threshold is retained as the identification threshold 2104 of the identification threshold information 2100 for each model.

This completes the processing (step 2411).

The risk value calculation unit 113 of Example 2 calculates the risk value using the risk model constructed by the processing of FIG. 24 above. Since the procedure is the same as that of the first embodiment, the explanation is omitted (see FIG. 14). The calculated risk values are retained, for example, as type 2 diabetes onset risk values 2202 to 2204 of the risk correction result information 2200 .

FIG. 25 is a flow chart showing an example of processing executed by the subject extraction unit 114 and the risk value correction unit 115 according to the second embodiment of the present invention.

When the process starts (step 2501), the subject extraction unit 114 sets extraction conditions (step 2502). This process is the same as step 1502 (FIG. 15) of the first embodiment.

Next, the subject extraction unit 114 reads the risk evaluation result (step 2503). Here, the risk of developing each disease of each person calculated by the risk value calculating unit 113 (for example, the type 2 diabetes developing risk value 2202 to the cerebrovascular disease developing risk value 2204 of the risk correction result information 2200) is read.

Next, the risk value correction unit 115 corrects the risk value read in step 2503 (step 2504). This correction is performed, for example, by the method described with reference to FIG. The corrected risk values are held as corrected risk values 2205 to 2207 for developing type 2 diabetes mellitus in the risk correction result information 2200, for example.

Next, the subject extraction unit 114 adds priorities for intervention such as health guidance based on the risk values corrected in step 2504 (step 2505). For example, when corrected risk values such as corrected onset risks 2314, 2324 and 2334 of subject extraction information 2300 are read, priority levels such as intervention priority levels 2315, 2325 and 2335 are added. At this time, the subject extracting unit 114 determines that the table 2310 includes only information about a person who is determined to have the highest risk of developing type 2 diabetes, based on the corrected risk of developing cardiovascular disease. Priority is given so that only information about persons determined to have the highest risk of developing cerebrovascular disease is included in table 2320 and only information about persons determined to have the highest risk of developing cerebrovascular disease is included in table 2330. You can choose a person.

Next, the target person extraction unit 114 stores the priority added in step 2305 as the target person extraction information 2300 (step 2506).

This completes the processing (step 2507).

Next, an example of a user interface provided by the risk analysis support system 101 of Example 2 will be described with reference to FIGS. 26 and 27. FIG.

FIG. 26 is an explanatory diagram showing an example of a user interface corresponding to the processing of the state determination unit 112 and the risk model construction unit 111 of Example 2 of the present invention.

A risk model building screen 2600 shown in FIG. 26 is an example of display data output by the risk analysis support system 101. For example, a data reading unit 2601, a model building condition setting unit 2602, a model building processing execution button 2603, and analysis data A set creation result display portion 2604 and a model construction result display portion 2605 are included. Except for the following differences, these are the data reading section 1601, the model construction condition setting section 1602, the model construction processing execution button 1603, the analysis data set creation result display section 1604, and the model construction of the risk model construction screen 1600 of the first embodiment. This is the same as the result display section 1605 . The differences are described below.

The analysis data set created in step 2407 is displayed in the analysis data set creation result display section 2604 . Specifically, for each person, attributes such as gender and age as explanatory variables, information such as BMI and interview results obtained from medical checkup results, etc., presence or absence of pre-existing diseases obtained from medical history information, and purpose Onset diseases, etc., which are variables, are displayed.

Information on the model constructed by the risk model construction unit 111 is displayed in the model construction result display unit 2605 . For example, information corresponding to the model ID 2101, the model name 2102, and the model parameter 2103 among the information included in the model identification threshold information 2100 may be displayed.

FIG. 27 is an explanatory diagram showing an example of a user interface corresponding to the processes of the risk value calculation unit 113, the subject extraction unit 114, and the risk value correction unit 115 according to the second embodiment of the present invention.

A target person extraction screen 1700 shown in FIG. 27 is an example of display data output by the risk analysis support system 101, and includes, for example, a data reading unit 2701, a target person extraction condition setting unit 2702, a target person extraction execution button 2703, and an object A person extraction result display portion 2704 is included. These are the same as the data reading portion 1701, the subject extraction condition setting portion 1702, the subject extraction execution button 1703, and the subject extraction result display portion 1704 of the subject extraction screen 1700 of the first embodiment, except for the following differences. be. The differences are described below.

The target person extraction result display unit 2704 displays the processing results of the risk value calculation unit 113, the target person extraction unit 114, and the risk value correction unit 115. For example, information corresponding to the target person extraction information 2300 may be displayed.

According to the second embodiment described above, it is possible to appropriately determine the priority of health guidance for preventing changes in health conditions such as the onset or aggravation of diseases. In particular, even when the frequency of occurrence differs depending on the mode of change in health condition, it is possible to appropriately compare the risks of changes in health condition and determine priority, and to overlook risks such as diseases with low occurrence frequency. is prevented.

Also, the system of the embodiment of the present invention may be configured as follows.

(1) A risk analysis support system, comprising a processor (for example, CPU 104) and a storage device (for example, memory 105, storage medium 106, and at least one of other storage media storing database 107) connected to the processor , and the storage device stores health information related to the health of a plurality of persons (for example, information managed by the medical examination history information management unit 120 and the health checkup information management unit 121) and attribute information of a plurality of persons (for example, attribute information managed by the information management unit 122) and a plurality of health condition definition information (for example, at least one of the severity definition information 700 and the target disease definition information 1900), and the processor stores the health information, attribute building a risk model for calculating the risk of a change in health condition based on the information and the defining information of multiple health conditions (e.g., step 1309 or step 2409); and based on the health information, the attribute information and the risk model, Calculate a risk value indicating the risk of change in health status of the plurality of persons (eg, step 1407), and calculate the priority of insurance guidance for the plurality of persons based on the risk value (eg, step 1504 or step 2505).

By doing this, it is possible to appropriately determine the priority of health guidance to prevent changes in health conditions such as the onset or aggravation of diseases.

(2) In (1) above, the plurality of health condition definition information includes information defining a plurality of severity levels of a particular disease (for example, severity definition information 700), and the risk model is such that each health condition is including a model (e.g., risk model parameter information 1000) for calculating the risk of changing from a severity level to a higher severity level, the processor, based on the health information and the plurality of health condition defining information, a plurality of Compute the person's current severity level (e.g., step 1206 and step 1207), and based on the health information, the demographic information, and the risk model, determine if the health status of the plurality of persons is from the current severity level to a higher severity level. is calculated as a risk value (e.g., step 1407), and for each current severity level of multiple persons, the health status is at high risk of changing from the current severity level to a higher severity level The priority of insurance guidance for a plurality of persons is calculated so that the higher the priority (for example, step 1504).

(3) In (1) above, the plurality of health condition definition information includes information that defines the onset of a plurality of diseases (for example, the target disease definition information 1900), and the risk model is the development of each of the plurality of diseases. A model for calculating risk (for example, a risk model included in the model-specific identification threshold information 2100) is included, and the processor, based on the health information, the attribute information, and the risk model, develops each of a plurality of diseases by a plurality of people. The risk of developing a disease is calculated as a risk value (for example, step 1407), and the priority of insurance guidance for a plurality of persons is calculated so that the higher the risk of developing a disease, the higher the priority for each disease (for example, step 2505). ).

(4) In (1) above, the storage device holds history information (for example, guidance history information 900) of health guidance given to a plurality of persons in the past, and the processor provides insurance guidance based on the history information. Exclude persons determined to be unnecessary from the calculation of insurance guidance priority.

(5) In (4) above, the processor determines that the same insurance guidance given more than a predetermined number of times in the past is unnecessary based on the history information.

(6) In (4) above, the history information includes information indicating whether treatment by the medical institution has been started for each of the plurality of persons, and the processor determines whether treatment by the medical institution has been started based on the history information. It is determined that insurance guidance for a person is unnecessary.

(7) In (1) above, the storage device holds risk values calculated in the past for a plurality of persons, and the processor identifies from the risk values calculated in the past and the newly determined risk values. Calculate the priority of insurance guidance for multiple persons based on the trend of change in risk values.

(8) In (1) above, the health information includes information indicating medical examination histories of a plurality of persons at medical institutions (for example, information managed by the medical examination history information management unit 120), and health examinations received by a plurality of persons. (for example, information managed by the health checkup information management unit 121) indicating the result of the health checkup.

(9) In (1) above, the risk model is a risk model (for example, the risk model included in the model-specific identification threshold information 2100) for calculating the risk of a change to the health condition for each health condition. The storage device stores, for each health condition, a threshold (for example, included in the model identification threshold information 2100) for determining whether a change to the health condition will occur based on the risk of a change to the health condition. 2104), the processor calculates, for each health condition, a risk value indicative of the risk of a change to that health condition occurring (e.g., step 1407), and for each health condition, based on the threshold: The risk values are corrected (for example, step 2504), and based on the corrected risk values, insurance guidance priorities for a plurality of persons are calculated (for example, step 2505).

(10) In (9) above, the processor corrects the risk value by at least one of lowering the risk value of a health condition with a high threshold and increasing the risk value of a health condition with a low threshold. do.

(11) In (9) above, the plurality of health condition definition information includes information defining the onset of a plurality of diseases (for example, the target disease definition information 1900), and the risk model develops each of the plurality of diseases. A model for calculating risk (for example, a risk model included in the model-specific identification threshold information 2100) is included, the threshold is a threshold for determining whether each of a plurality of diseases develops, and the processor is Based on the health information, attribute information, and risk model, the risk of each of the plurality of persons developing each of the plurality of diseases is calculated as a risk value (for example, step 1407), and the corrected risk value is calculated for each of the plurality of persons. identified the disease with the highest risk of developing the disease as the disease with the highest risk of developing the disease, and for each disease, multiple individuals were analyzed so that among those with the highest risk of developing the disease, the higher the adjusted risk value, the higher the ranking. Calculate the priority of insurance guidance for (eg, step 2505).

It should be noted that the present invention is not limited to the above-described embodiments, and includes various modifications. For example, the above embodiments have been described in detail for better understanding of the present invention, and are not necessarily limited to those having all the configurations described. Also, it is possible to replace part of the configuration of one embodiment with the configuration of another embodiment, or to add the configuration of another embodiment to the configuration of one embodiment. Moreover, it is possible to add, delete, or replace a part of the configuration of each embodiment with another configuration.

In addition, each of the above configurations, functions, processing units, processing means, etc. may be implemented in hardware by designing, for example, integrated circuits in part or in whole. Moreover, each of the above configurations, functions, etc. may be realized by software by a processor interpreting and executing a program for realizing each function. Information such as programs, tables, files, etc. that realize each function is stored in storage devices such as non-volatile semiconductor memories, hard disk drives, SSDs (Solid State Drives), or computer-readable non-storage devices such as IC cards, SD cards, DVDs, etc. It can be stored on a temporary data storage medium.

In addition, the control lines and information lines indicate what is considered necessary for explanation, and not all control lines and information lines are necessarily indicated on the product. In fact, it may be considered that almost all configurations are interconnected.

Claims (12)

1. A risk analysis support system,
   a processor and a storage device connected to the processor;
   the storage device holds health information related to the health of a plurality of persons, attribute information of the plurality of persons, and definition information of a plurality of health conditions;
   The processor
   constructing a risk model for calculating the risk of a change in the health condition based on the health information, the attribute information, and the plurality of health condition definition information;
   calculating a risk value indicating a risk of change in the health condition of the plurality of persons based on the health information, the attribute information, and the risk model;
   A risk analysis support system, wherein priority of insurance guidance for said plurality of persons is calculated based on said risk values.
2. The risk analysis support system according to claim 1,
   the plurality of health condition definition information includes information defining a plurality of severity levels of a particular disease;
   the risk model includes a model for calculating the risk of the health condition changing from each severity level to a higher severity level;
   The processor
   calculating current severity levels of the plurality of persons based on the health information and the plurality of health condition definition information;
   calculating, as the risk value, a risk that the health condition of the plurality of persons will change from a current severity level to a higher severity level based on the health information, the attribute information, and the risk model;
   Insurance guidance to the plurality of persons such that, for each current severity level of the plurality of persons, the higher the risk of the health condition changing from the current severity level to a higher severity level, the higher the ranking. A risk analysis support system characterized by calculating the priority of
3. The risk analysis support system according to claim 1,
   The definition information of the plurality of health conditions includes information defining the onset of a plurality of diseases,
   the risk model includes a model for calculating the risk of developing each of the plurality of diseases;
   The processor
   calculating, as the risk value, the risk of the plurality of persons developing each of the plurality of diseases based on the health information, the attribute information, and the risk model;
   A risk analysis support system, wherein, for each disease, a priority order of insurance guidance for the plurality of persons is calculated such that the higher the risk of developing the disease, the higher the priority.
4. The risk analysis support system according to claim 1,
   The storage device holds history information of health guidance given to the plurality of persons in the past,
   The risk analysis support system, wherein the processor excludes the person for whom the insurance guidance is determined to be unnecessary based on the history information from the calculation of the priority of the insurance guidance.
5. The risk analysis support system according to claim 4,
   The risk analysis support system, wherein the processor determines, based on the history information, that the same insurance guidance given a predetermined number of times or more in the past is unnecessary.
6. The risk analysis support system according to claim 4,
   The history information includes information indicating whether treatment by a medical institution has been started for each of the plurality of persons,
   The risk analysis support system, wherein the processor determines, based on the history information, that the insurance guidance for a person for whom treatment by a medical institution has been started is unnecessary.
7. The risk analysis support system according to claim 1,
   the storage device holds the risk values calculated in the past for the plurality of persons;
   The processor calculates a priority order of insurance guidance for the plurality of persons based on trends in changes in the risk values identified from the previously calculated risk values and the newly calculated risk values. A risk analysis support system characterized by:
8. The risk analysis support system according to claim 1,
   A risk analysis support system, wherein the health information includes at least one of information indicating medical institution visit histories of the plurality of persons and information indicating results of health examinations received by the plurality of persons.
9. The risk analysis support system according to claim 1,
   The risk model includes, for each health condition, a risk model for calculating the risk of a change to the health condition,
   The storage device holds, for each health condition, a threshold for determining whether or not a change to the health condition will occur based on the risk of a change to the health condition,
   The processor
   calculating, for each health condition, a risk value indicating the risk of a change to the health condition;
   correcting the risk value based on the threshold value for each health condition;
   A risk analysis support system, wherein priority of insurance guidance for the plurality of persons is calculated based on the corrected risk values.
10. A risk analysis support system according to claim 9,
    The processor corrects the risk value by at least one of lowering the risk value for the health condition for which the threshold is high and increasing the risk value for the health condition for which the threshold is low. A risk analysis support system characterized by:
11. A risk analysis support system according to claim 9,
    The definition information of the plurality of health conditions includes information defining the onset of a plurality of diseases,
    the risk model includes a model for calculating the risk of developing each of the plurality of diseases;
    The threshold is a threshold for determining whether each of the plurality of diseases develops,
    The processor
    calculating, as the risk value, the risk of the plurality of persons developing each of the plurality of diseases based on the health information, the attribute information, and the risk model;
    For each of the plurality of persons, the disease with the highest risk value after the correction is specified as the disease with the highest risk of developing;
    calculating the priority of insurance guidance for each of the plurality of persons such that the higher the risk value after correction, the higher the rank among persons having the highest risk of developing the disease for each of the diseases; A risk analysis support system.
12. A risk analysis support method executed by a computer system,
    The computer system has a processor and a storage device connected to the processor,
    the storage device holds health information related to the health of a plurality of persons, attribute information of the plurality of persons, and definition information of a plurality of health conditions;
    The risk analysis support method includes:
    a step in which the processor constructs a risk model for calculating the risk of the health condition changing, based on the health information, the attribute information, and the plurality of health condition definition information;
    a step of calculating a risk value indicating the risk of change in the health condition of the plurality of persons based on the health information, the attribute information and the risk model;
    and calculating a priority order of insurance guidance for the plurality of persons based on the risk values.

Images