JP2007334781A - Health instruction support system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a health instruction support system capable of setting a threshold value conformed with information of an instructor to create risk knowledge. <P>SOLUTION: The health instruction support system includes a risk knowledge preparing means 106 for calculating an onset rate with respect to a combination of classifications of health examination items classified by an initial threshold value, and a support degree indicating reliability of the onset rate, and for creating the risk knowledge, a contradiction generation number calculating means 109 for extracting a rule with a change of the onset rate brought into a contradictory relation with respect to the risk knowledge, and for calculating a contradiction generation number, and a support degree extracting means 126 for extracting a support degree of the classification classified by the threshold of the health examination item generating the contradiction, based on the risk knowledge. The health instruction support system further includes a threshold change classification selecting means 113 for selecting the classification of a low support degree under the condition of the item where the contradiction included most in a condition of the rule is generated, based on the rules brought into the contradictory relation and the support degree, as a threshold changing classification, and a threshold changing means 114 for changing the threshold with the contradiction generating number getting lowest when changed, based on the support degree and the contradiction generating number, toward a direction with the support degree in the selected classification getting high. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a health guidance support system that presents information for disease prevention / health promotion from health checkup results.

  There is a health guidance support system that analyzes health checkup information accumulated at health checkup facilities, predicts diseases that may develop in the future based on the results, and supports health guidance. For example, in Patent Document 1, a threshold is set for each medical examination item, and the incidence rate that indicates the percentage of people who will have future test abnormalities by test value rank, age group, lifestyle pattern, and family history divided by the threshold. To calculate risk knowledge. And the system which calculates | requires the onset rate applicable to the medical examination result of the guidance subject from the risk knowledge, and is instructing is introduced. In such a system, for example, when the threshold value of blood glucose level is set to 110 and the threshold value of HbA1c, which is one of the indicators of diabetes, is set to 5.6, the blood glucose level is less than 110 and less than HbA1c 5.6 If you enter, risk knowledge suggests a 20% diabetes incidence with a blood glucose level below 110 and HbA1c5.6. Also, if you enter a blood glucose level of less than 110 and HbA1c5.6 or higher, you will be presented with a diabetes incidence of 30% with a blood glucose level of less than 110 and HbA1c5.6 or higher. I do.

JP 2002-183647 A

  In such a system that provides health guidance by presenting risks based on the analysis results of the medical examination information, it is necessary to create risk knowledge that presents guidance content that is easy for a leader who provides health guidance. Risk knowledge that is easy for the instructor to instruct is based on the knowledge of the instructor when the threshold for dividing the value of the health check item is in line with the standard value of academic societies, etc., and when the test value changes beyond the above threshold. It is necessary that the values change in agreement. However, in such a system, depending on the threshold value of the health examination items, an incidence rate that contradicts the knowledge of the instructor may be presented. For example, if the threshold setting is poor, the blood glucose threshold is 110, and the HbA1c threshold is 5.8, the blood sugar level is less than 110, HbA1c5 is less than 110, HbA1c5 Diabetes incidence of 23% or more may be presented. In this case, the higher the value of HbA1c, the lower the incidence, which contradicts the knowledge of the leader and makes it difficult to guide. However, in the above conventional example, there is no specific description about a risk knowledge creation method that is easy for such an instructor to teach.

  In addition, one of the causes of the contradiction as described above may be that the reliability of the presented onset rate is low because the number of affected individuals is divided depending on the threshold value. In this case, as a solution method, a method is conceivable in which a threshold is set so that there are sufficient onsets in a category with few onsets. For example, a method of introducing a support level indicating the value obtained by dividing the number of patients in the segment divided by the threshold value by the number of populations, and setting the threshold value so that the support level of the low support level becomes high is considered. It is done. This method sets the threshold by changing the threshold with the higher support level out of the two thresholds when the low support level is divided by two thresholds. . However, a specific threshold setting method has not been taken into consideration, for example, the changed threshold value may not always be the threshold value with the least inconsistency.

  An object of the present invention is to provide a health guidance support system that solves the above-described problems and creates risk knowledge by setting a threshold value with less contradiction in the vicinity of a reference value of an academic society or the like.

  In order to solve the above-mentioned problems and realize the object, the health guidance support system of the present invention includes an initial threshold setting means for setting an initial threshold value for classifying items of medical checkup, and a medical examination divided by the set threshold value. Based on the category of the item, the combination of the conditions and the incidence rate indicating the ratio of the number of affected individuals to the combination, and the value of the number of affected individuals divided by the number of population, the support level indicating the reliability of the incidence rate The risk knowledge creation means for calculating risk and creating risk knowledge, and for the created risk knowledge, extract the rule whose change in the incidence is inconsistent with the change in the condition of the health checkup item, A contradiction occurrence number calculating means for calculating the number of combinations as a contradiction occurrence number, a support degree extracting means for extracting the support degree of the section divided by the threshold of the medical examination item in which the contradiction occurs from the created risk knowledge, and a contradiction In relation Threshold change category selection means for selecting, as a category for changing the threshold, a threshold change category selection means that selects a category with a low level of support in the condition of a medical examination item that contains the most contradiction in the rule condition and the support level. And a threshold value changing means for changing the threshold value at which the number of contradiction occurrences becomes the smallest when it is changed in a direction in which the degree of support of the selected category is increased.

  Further, the health guidance support system of the present invention has a high contribution as an item for changing the threshold value from the contribution degree and the number of occurrences of contradiction, and the disease contribution degree calculation means for calculating the contribution degree of the medical examination item to the disease. , And a threshold change item selection means for selecting a medical examination item in which a contradiction occurs.

  Furthermore, the health guidance support system of the present invention is characterized by having contradiction occurrence information display means for displaying a list of contradiction occurrence numbers calculated by the contradiction occurrence number calculation means.

  Furthermore, the health guidance support system according to the present invention corresponds to the health check result from the health check result input means for inputting the health check result of the target person and the risk knowledge created by changing the threshold value by the threshold change means. The information acquisition means which acquires the rule to perform, etc. and the guidance content display means which displays the onset rate of the rule acquired by the information acquisition means are characterized.

  In the health guidance support system of the present invention, the threshold value change category selection means is a category in which the threshold value is changed based on the contradictory rules calculated by the contradiction occurrence number calculation means and the support level extracted by the support level extraction means. The threshold value changing means changes the threshold value of the above classification from the initial threshold value and sets it while confirming the number of contradictions and the support level. As a result, it is possible to set a threshold value that has a high degree of support in the vicinity of the initial threshold value and that has less contradiction, and has the effect of being able to create risk knowledge that presents guidance content that is easy for a leader to teach.

  Furthermore, in the health guidance support system of the present invention, the threshold change item selection means calculates the number of contradictions from the disease contribution degree information calculated by the disease contribution degree calculation means in the order of items having the highest contribution to the disease. Select an item whose number of occurrences is not 0 as a threshold change item. Thereby, there is an effect that the contradiction can be reduced by changing the threshold in order from an important item having a high contribution to the disease.

  Furthermore, in the health guidance support system of the present invention, the contradiction occurrence information display means displays a list of contradiction occurrence number information and allows the operator to determine a final threshold value. As a result, a list of the number of contradiction occurrences from the initial threshold value to the optimum threshold value is displayed, so that the operator can grasp the tendency of the contradiction occurrence number due to the threshold setting, and the number of contradiction occurrences is small in the vicinity of the initial threshold that is easy to guide. There is an effect that a threshold value can be selected.

  Furthermore, in the health guidance support system of the present invention, the contradiction occurrence number calculation means extracts only the rules used for guidance from the risk knowledge candidates in order to reduce the number of rule searches accompanying the calculation of the contradiction occurrence number, The number of inconsistencies is calculated by checking for inconsistencies between usage rules. As a result, the number of rule searches can be reduced, and the number of contradictions can be calculated at high speed.

  Furthermore, in the health guidance support system of the present invention, the health examination result input means inputs the health examination result of the person being instructed, and the information acquisition means is a rule corresponding to the health examination result inputted from the risk knowledge for guidance. The guidance content display means displays the onset rate of the guidance target person. As a result, the instructing operator is able to provide effective health guidance by presenting the onset rate corresponding to the health check result of the instructed person from risk knowledge for guidance with less contradiction. .

  Hereinafter, the best mode for carrying out the present invention will be described in detail with reference to the drawings. In the following explanation, blood glucose levels and HbA1c are taken as examples of health check items for changing the threshold, and diabetes is taken as an example of the disease, and the risk knowledge is constructed by setting two threshold values for both items.

  FIG. 1 is a diagram illustrating a configuration example of a health guidance support system according to an embodiment of the present invention. The health guidance support system includes a health guidance support terminal 101 and a database 102. The health guidance support terminal 101 is a computer device, and includes an input device 103 such as a mouse and a keyboard, an output device 104 such as a display and a printer, a CPU 105 that calculates and executes a program, and a recording device 106 such as a hard disk and a memory. . The storage device 106 stores means consisting of programs and data described below, which are read out and executed by the CPU 105 as necessary.

  The storage device 106 uses the initial threshold setting means 107 for setting the initial threshold of the medical examination items used for instruction, and the correlation rule mining to comprehensively display the incidence and support for each medical examination item divided by the threshold. From the risk knowledge creating means 108 for creating risk knowledge, the disease contribution calculating means 110 for calculating the contribution to the disease of the test value / interview result using a statistical method, and the disease contribution information, Threshold change item selection means 111 for selecting a medical examination item having a high contribution to a disease as a threshold change item, and a contradiction occurrence number for calculating a contradiction occurrence number of the medical examination item selected by the threshold change item selection means 111 Calculation means 109, information acquisition means 112 for acquiring medical examination information, risk knowledge candidates, contradiction occurrence number information, contradiction detailed information, support level information, disease contribution level information, rules, etc. from database 102, contradiction detailed information, support Degree information A threshold changing section selecting means 113 for selecting a section for changing the threshold, a threshold changing means 114 for determining the direction and width of the selected threshold from the inconsistency occurrence number information and the support level information, and changing the threshold; Contradiction information display means 115 for displaying a history of the number of contradictions from the initial threshold to the optimum threshold and selecting the threshold, medical examination result input means 116 for inputting the medical examination result of the examinee, and the entered medical examination A guidance content display means 117 for extracting the onset rate corresponding to the result from the guidance risk knowledge and a support degree extraction means 126 for extracting support degree information from the risk knowledge candidates are stored. The database 102 includes a medical examination information management unit 118 that manages medical examination information, a threshold information management unit 119 that manages threshold information, a risk knowledge candidate management unit 120 that manages risk knowledge candidates, and contradiction occurrence number information. , Contradiction occurrence information management means 121 for managing contradiction detailed information, support degree information management means 125 for managing support degree information, disease contribution degree information management means 122 for managing disease contribution degree information, and guidance used for guidance A risk knowledge management means 124 for guidance for managing risk knowledge is provided.

  FIG. 2 is a diagram illustrating an example of the medical examination information managed by the medical examination information management unit 118. In addition to personal ID 201 that identifies the individual, consultation date 202, age 203 at the time of the medical examination, etc., as a test value, blood glucose level 204, HbA1c205, and BMI206 test results determined by the doctor as a determination result for diabetes 207 It manages information such as.

  FIG. 3 is a diagram showing an example of the initial threshold information managed by the threshold information management means 119. The medical examination item 301 and initial threshold values 302 to 303 of the item are managed. For example, two age thresholds of 40 and 50 indicate that the age is under 40 years old, 40 to 49 years old, and 50 years old or older.

  FIG. 4 is a diagram showing an example of disease contribution information managed by the disease contribution information management means 122. As shown in FIG. The medical examination item 401 and the contribution degree 402 to diabetes of the item are managed. In this case, the contribution value 402 indicates that the contribution is higher as the numerical value of the contribution degree 402 is younger, and the contribution degree 1 item indicates the highest contribution.

  FIG. 5 is a diagram showing an example of risk knowledge candidates managed by the risk knowledge candidate management means 120. As shown in FIG. The rules 508 to 518 indicating the incidence 507 of diabetes and the support 506 of a person having a combination of a plurality of conditions such as age 502, BMI 503, blood glucose level 504, HbA1c505, etc., which the condition unit 501 has are managed. Here, the onset rate 507 indicates the number of onset persons in the group having the same test value / interview result condition divided by the number of persons in the group. Support level 506 is obtained by dividing the number of patients in the group having the same test value / interview result condition by the number of the population, and indicates the reliability of the incidence.

  FIG. 6 is a diagram showing an example of the contradiction occurrence number information managed by the contradiction occurrence information management unit 121. Threshold candidate ID (601) for identifying a blood sugar level and a threshold value for HbA1c, a blood sugar level threshold A (602), a blood sugar level threshold B (603), an HbA1c threshold value A (604), an HbA1c threshold value B (605) ), Blood glucose level threshold A (602), blood glucose level threshold B (603), HbA1c threshold A (604), HbA1c threshold B (605), the number of inconsistent blood glucose levels 606, HbA1c The number of contradictions 607 is managed. For example, 608 is explained. The value 1 of the threshold candidate ID (601) is an ID indicating the blood sugar level thresholds 100 and 110, and the HbA1c thresholds 5.6 and 5.8. The number of contradictions with this ID is the blood sugar level. The number of inconsistencies when the threshold values 100 and 110 of HbA1c and the threshold values 5.6 and 5.8 of HbA1c are shown. In addition, the blood glucose level threshold A (602) is 100, the blood glucose level first threshold is 100, the blood glucose level threshold B (603) is 110, the blood glucose level second threshold is 110, and the HbA1c threshold is A. A value 5.6 of (604) indicates that the first threshold value of HbA1c is 5.6, and a value 5.8 of threshold value B (605) of HbA1c indicates that the second threshold value of HbA1c is 5.8. In addition, the value 0 of the blood sugar level contradiction occurrence number 606 is the value of the blood sugar level contradiction occurrence number 0 when the blood glucose level threshold is 100, 110, the threshold value 5.6, 5.8 of the HbA1c, and the value 607 of the HbA1c contradiction occurrence number. 3 indicates that the number of contradictions in HbA1c is 3 when the blood glucose threshold values are 100 and 110, and the HbA1c threshold values are 5.6 and 5.8.

  FIG. 7 is a diagram showing an example of support level information managed by the support level information management means 125. As shown in FIG. A threshold candidate ID (601) for identifying a blood sugar level and a threshold candidate for HbA1c and nine support levels 801 to 809 are managed. Here, the nine support levels 801 to 809 are the blood glucose threshold A (602), the blood glucose threshold B (603), the HbA1c threshold A (604), and the HbA1c threshold B (605) in FIG. It shows the support level of the rules of the 9 sections. The support levels A to C (801 to 803) are the support levels of the rule that the blood glucose level is less than the threshold A (602) and the HbA1c is less than the threshold A (604), and the blood glucose level is less than the threshold A (602) and the HbA1c is the threshold A. (604) The support level of a rule that is less than or equal to the threshold value B (605), and the support level of a rule that has a blood glucose level less than the threshold value A (602) and HbA1c is equal to or greater than the threshold value B (605).

  In addition, the support levels D to F (804 to 806) are the support levels of the rules in which the blood glucose level is greater than or equal to the threshold A (602) and less than the threshold B (603) and the HbA1c is less than the threshold A (604). 602) Support for rules with HbA1c greater than or equal to threshold B (603) and HbA1c greater than or equal to threshold A (604) and less than threshold B (605), blood glucose level greater than or equal to threshold A (602) and less than threshold B (603), HbA1c is threshold B (605) The degree of support for the above rules is shown.

  Furthermore, the support levels G to I (807 to 809) are the support levels of the rules in which the blood glucose level is equal to or higher than the threshold value B (603) and the HbA1c is lower than the threshold value A (604), and the HbA1c is equal to or higher than the threshold value B (603). The support level of a rule with a threshold value A (604) or more and less than a threshold value B (605), and the support level of a rule with a blood glucose level of a threshold value B (603) or more and HbA1c of a threshold value B (605) or more.

  For example, 810 will be described. The value 1 of the threshold candidate ID 601 is an ID indicating the blood glucose level thresholds 100 and 110, the HbA1c thresholds 5.6 and 5.8, and the support with this ID is the blood glucose level threshold 100, The degree of support for 110 and HbA1c thresholds 5.6 and 5.8 is shown. A support level A (801) value of 0.3% is the support level of a rule whose blood glucose level is less than 100 and HbA1c is less than 5.6, and support level B (802) value of 0.05% is a blood glucose level of less than 100 and HbA1c is 5.6 or more The support level of a rule of less than 5.8, the support level C (803) value of 0.8% indicates the support level of a rule having a blood glucose level of less than 100 and HbA1c of 5.8 or more. In addition, a support level D (804) value of 0.3% is a support level of a rule whose blood glucose level is 100 or more and less than 110 and HbA1c is less than 5.6, and a support level E (805) value of 0.15% is a blood glucose level of 100 or more and 110 The support level of a rule with HbA1c of less than 5.6 and less than 5.8 and a support level F (806) value of 0.9% indicates the support level of a rule with a blood glucose level of 100 to less than 110 and HbA1c of 5.8 or more. In addition, the support level G (807) value 0.5% is the support level of the rule whose blood glucose level is 110 or higher and HbA1c is less than 5.6, and the support level H (808) value 0.2% is the blood glucose level 110 or higher and HbA1c is The support level of the rule of 5.6 or more and less than 5.8, and the support level I (809) value of 0.8% indicates the support level of the rule that the blood glucose level is 110 or more and HbA1c is 5.8 or more.

  FIG. 8 is a diagram showing an example of detailed contradiction information managed by the contradiction occurrence information management unit 121. Threshold candidate ID (601) for identifying blood sugar level and HbA1c threshold candidates, contradictory detail ID (702) for identifying contradictory rules, and the onset of diabetes in a person who has a combination of rule conditions 502 to 505 and the conditions It manages rate 507 and support 506. In this case, the value 1 of the threshold candidate ID (601) is an ID indicating the blood glucose level thresholds 100 and 110, and the HbA1c threshold values 5.6 and 5.8. This rule is inconsistent when the threshold values of HbA1c are 5.6 and 5.8. Further, the value 1 of the contradiction detail ID (702) indicates that the rule 709 and the rule 710 are rules having a contradiction relationship, and the rule 709 and the rule 710 ( The contradiction detail ID value 1), the rules 711 and 712 (the contradiction detail ID value 2), and the rules 713 and 714 (the contradiction detail ID value 3) are inconsistent rules.

  FIG. 9 is a diagram illustrating an example of the support level of the division rule divided by the blood glucose level and the threshold value of HbA1c. In the example in this figure, the threshold values for blood glucose levels A: 100 (1101), B: 110 (1102), HbA1c threshold values A: 5.6 (1103), B: 5.8 (1104) The support degrees 1105 to 1113 are shown, corresponding to the support degrees A to I (801 to 809) of 810 in FIG.

  FIG. 10 shows a contradiction occurrence number display screen in which the contradiction occurrence information display means 115 displays a list of the blood glucose level from the initial threshold value to the optimum threshold value and the number of contradiction occurrences of HbA1c on the output device 104 and allows the operator to select a threshold value. It is a figure which shows an example of 1200. Threshold values 1201 to 1204 indicate blood sugar level threshold values, and threshold values 1205 to 1209 indicate HbA1c threshold values. The contradiction occurrence number buttons 1210 to 1214 display the blood glucose level contradiction occurrence number and the HbA1c contradiction occurrence number when the blood glucose level threshold and the HbA1c threshold are set on the buttons. In this case, the contradiction occurrence number buttons 1212 and 1213 indicate the number of occurrences of contradiction when the initial threshold of blood glucose level is set to 100 (1203), 110 (1202), and the initial threshold of HbA1c is set to 5.6 (1207), 5.8 (1208). Are displayed on the button as the number of contradictions in blood glucose level 0 (1218, 1219) and the number of contradictions in HbA1c 3 (1223, 1224).

  Furthermore, the contradiction occurrence number button 1214 does not change the blood glucose thresholds 100 (1203), 110 (1202), and the HbA1c threshold 5.6 (1207) from the initial threshold, and changes the HbA1c threshold 5.8 (1208) to 5.9 (1209). The number of contradiction occurrences when changed to is displayed on the button as the number of contradiction occurrences of blood glucose level 0 (1220) and the number of contradiction occurrences of HbA1c 2 (1225).

  In addition, the contradiction occurrence number button 1211 does not change the blood glucose thresholds 100 (1203), 110 (1202), the HbA1c threshold 5.8 (1207) from the initial threshold, and the HbA1c threshold 5.6 (1207) 5.5 (1206) On the button, the number of contradictory occurrences 0 (1217) and the number of contradictory occurrences 1 (1222) of HbA1c are displayed on the button.

  To determine the threshold value, first, either the contradiction occurrence button 1212 or 1213 is pressed, and if 1213 is pressed, from among the contradiction occurrence buttons 1210 to 1212, if 1212 is pressed, the contradiction occurrence button 1213 This is done by selecting from ˜1214 and finally pressing the threshold value decision button 1215. For example, first, when the contradiction occurrence number button 1213 is pressed, then 1210 is selected and the threshold value decision button 1215 is pressed, the blood glucose level threshold is 100 (1203), 110 (1202), and the threshold value of HbA1c is 5.4 (1205) , 5.8 (1208), which is the optimal threshold value where the number of contradictions between the blood glucose level and HbA1c is 0.

  FIG. 11 is a diagram showing an example of an initial threshold value input screen 1301 that the initial threshold value setting means 107 displays on the output device 104 and inputs the initial threshold value of the medical examination item. The first threshold value input columns 1302 to 1305 indicate columns for inputting the first threshold value of the medical examination item. The second threshold value input columns 1306 to 1309 indicate columns for inputting the second threshold value of the medical examination item. Reference numeral 1310 denotes a determination button. For example, if 100 is entered in the first threshold value entry field 1303 of the blood glucose level, 110 is entered in the second threshold value entry field 1307 and the enter button 1310 is pressed, the blood sugar level is divided into three categories of less than 100, 100 to 110, 110 or more. It is done. The initial threshold information of the medical examination items input on this screen is managed in the format shown in FIG.

  FIG. 12 shows a risk in which only the rules used for guidance are extracted from the risk knowledge candidates in FIG. 5 managed by the risk knowledge candidate management means 120 in order for the contradiction occurrence number calculating means 109 to calculate the number of contradictory occurrences of the medical examination items. It is a figure which shows an example of a knowledge candidate. As the rules used for guidance, conditions such as age 502, BMI503, blood glucose level 504, HbA1c505, support degree 507, and incidence 506 are shown.

  FIG. 13 shows the inconsistency relation extracted by checking the contradiction between rules for the risk knowledge candidates of only the usage rules in FIG. 12 in order for the contradiction occurrence number calculating means 109 to calculate the inconsistency occurrence number of the medical examination items. It is a figure which shows an example of a certain rule. In this example, when the rule 1507 including the condition less than HbA1c5.6 is extracted from the risk knowledge candidates of only the usage rules in FIG. 12, and only the condition of HbA1c in this rule is changed from less than 5.6 to less than 5.6 and less than 5.8 The rules that are inconsistent are extracted. 1508 is a condition for extracting a rule used for guidance from the risk knowledge candidates of only the usage rule in FIG. 12 when only the HbA1c condition of the rule 510 is changed from less than 5.6 to less than 5.6 and less than 5.8. In this case, the condition of HbA1c is 5.6 to 5.8, and the other conditions are the same as rule 510 or null. Rules 517 and 511 indicate the rules extracted under the extraction condition 1508. The rule 517 having a higher incidence than the rule 510 is a rule having no contradiction, and the rule 511 having a lower incidence than the rule 510 is The rules are in contradiction.

  FIG. 14 is a diagram showing an example of guidance risk knowledge managed by the guidance risk knowledge management means 124. The rules 1502 to 1505 indicating the incidence 507 of diabetes and support 506 of a person having a combination of a plurality of conditions such as age 502, BMI 503, blood glucose level 504, HbA1c505, etc. possessed by the condition unit 501 are managed.

  Next, the operation will be described in detail using a flowchart and a sequence diagram. First, an example of a procedure for creating guidance risk knowledge from medical examination information will be described with reference to the flowcharts of FIGS. 15 to 18, the flowchart of FIG. 23, and the sequence diagram of FIG. FIG. 15 is a diagram illustrating an example of a flowchart showing a flow of processing for creating risk knowledge for guidance from medical examination information. FIG. 16 is a diagram showing an example of a detailed flowchart of the change threshold determination step 1710 for determining the threshold to be changed in the flowchart of FIG. FIG. 17 is a diagram showing an example of a detailed flowchart of the contradiction occurrence number calculating step 1706 for calculating the contradiction occurrence number of the medical examination items in the flowcharts of FIGS. 15 and 16. FIG. 18 is a diagram showing an example of a detailed flowchart of the threshold change item selection step 1705 for selecting a medical examination item for changing the threshold in the flowchart of FIG. FIG. 23 is a diagram showing an example of a detailed flowchart of the threshold search step 1709 for searching for a threshold with less contradiction in the flowchart of FIG. FIG. 19 is an example of a sequence diagram showing exchange between the health guidance support terminal 101 and the database 102 in the flowchart of FIG.

  When the creation of guidance risk knowledge is started (1701), an initial threshold value input step 1702 is first performed. Here, the initial threshold value setting means 107 displays the initial threshold value input screen 1301 of FIG. 11 on the output device 104 to input the initial threshold value of the medical examination items that are easy for the operator to guide. The operator uses the input device 103 to input an initial threshold value based on academic society standard values, the incidence distribution of single items of medical examination items, support distribution, etc., and the first threshold value input fields 1302 to 1305 and the second threshold value input. Fill in fields 1306-1309. In this case, the first threshold 40 (1302), the second threshold 50 (1306), the first threshold 100 (1303) of the blood glucose level, the second threshold 110 (1307), the first threshold 5.6 (1304) of HbA1c, The second threshold value 5.8 (1308), the first BMI threshold value 25 (1305), and the second threshold value 28 (1309) are input. When the determination button 1310 is pressed after the input is completed, the input initial threshold information is managed in the format shown in FIG. In the sequence diagram of FIG. 19, initial threshold information registration 2102 is performed in the database 102 from the health guidance support terminal 101.

  Next, a risk knowledge candidate creation step 1703 in FIG. 15 is performed. Here, first, the information acquisition unit 112 acquires the medical examination information of FIG. 2 managed by the medical examination information management unit 118 and the initial threshold information of FIG. 3 managed by the threshold information management unit 119. Next, the risk knowledge creating means 108 divides the medical examination information using the initial threshold information. Then, analysis by association rule mining is performed on the divided medical examination information, and a risk knowledge candidate indicating the onset rate and the support degree for each condition part is created by combining the values of the divided medical examination items. To do. Here, the diabetes incidence rate is obtained by extracting people who are not diabetic at the first time from the health examination information for a plurality of years, and obtaining the ratio of those who subsequently developed diabetes. The onset of diabetes is determined, for example, based on information determined by a doctor or when the fasting blood glucose level is 126 or higher. Further, as shown in FIG. 5, the prepared risk knowledge candidate is data in which a condition part 501 combining a plurality of conditions and a diabetes incidence 507 and a support degree 506 of a person having the condition part are recorded. For example, rule 508 has an 11% incidence rate for health checkup results of ages 40-49, BMI 25-28, blood glucose level 100-110, and HbA1c 5.6, and the degree of support that indicates the reliability of the rate. Indicates 0.1%. The risk knowledge candidate prepares a rule having a combination of such various conditions. In the sequence diagram of FIG. 19, the health guidance support terminal 101 makes a medical examination information and initial threshold information acquisition request 2103, acquires the medical examination information and initial threshold information 2104 from the database 102, and performs risk knowledge candidate registration 2105. .

  Next, a disease contribution calculation step 1704 is performed. Here, first, the information acquisition unit 112 acquires the medical examination information of FIG. 2 managed by the medical examination information management unit 118. Next, the disease contribution degree calculating means 109 analyzes the acquired medical examination information by logistic regression analysis, and calculates the contribution degree of the medical examination items to the onset of diabetes. The calculated disease contribution degree information is data in which the medical examination item 401 and the contribution degree 402 are recorded as shown in FIG. In this example, an item having a higher contribution to the disease is shown as the numerical value is younger, and a blood glucose level having a contribution of 1 is the item having the highest contribution.

  Next, a threshold change item selection step 1705 is performed. Here, the threshold change item selection unit 111 calculates the number of contradiction occurrences in order of items having a high contribution to the disease, and selects an item whose contradiction occurrence number is not 0 as the threshold change item. Thereby, it is possible to reduce contradiction by changing the threshold in order from an important item having a high contribution to the disease. An example of a specific procedure will be described using the flowchart of FIG. When selection of threshold change items is started (2401), first, a high contribution item selection step 2402 is performed. Here, first, the information acquisition unit 112 acquires the disease contribution level information of FIG. 4 managed by the disease contribution level information management unit 122. Next, the item with the highest contribution to the onset of diabetes is selected. In this case, a blood glucose level with a contribution of 1 is selected.

  Next, the contradiction occurrence number calculation step 1706 of FIG. 18 is performed. Here, first, the information acquisition unit 112 acquires the risk knowledge candidates of FIG. 5 managed by the risk knowledge candidate management unit 120. Next, the contradiction occurrence number calculating means 109 calculates the contradiction occurrence number of the blood glucose level selected in the high contribution item selection step 2402 for the risk knowledge candidate. Here, the number of occurrences of contradiction in blood glucose level means that if the condition of items other than blood glucose level is not changed and only the condition of blood glucose level is changed, the incidence of the higher blood glucose level becomes lower and the incidence is lower. The number of rule combinations to be reversed is calculated. A specific calculation procedure will be described later. In this case, it is assumed that the number of contradictory blood glucose levels is zero.

  Next, step 2403 for determining the number of contradictions is performed. Here, it is determined whether or not the contradiction occurrence number of the medical examination items calculated in the contradiction occurrence number calculation step 1706 is zero. In the case of 0, the process returns to the high contribution item selection step 2402. If it is not 0, the medical examination item is selected as a threshold change item, and selection of the threshold change item is terminated (2404). In this case, since the number of inconsistencies in the blood glucose level is 0, the process returns to the high contribution item selection step 2402. Then, HbA1c with contribution 2 is selected as the next highest contribution item from the disease contribution information in FIG. 4 managed by the disease contribution information management means 122. Next, in a contradiction occurrence number calculation step 1706, the contradiction occurrence number calculation means 109 calculates the contradiction occurrence number of HbA1c.

  Here, with respect to an example of a specific procedure for calculating the number of contradiction occurrences, a detailed flowchart of the contradiction occurrence number calculation step 1706 in FIG. 17, risk knowledge candidates in FIG. 5, risk knowledge candidates only in use rules in FIG. 12, FIG. Will be described. The calculation item is HbA1c. The risk knowledge candidates shown in FIG. 5 created by the risk knowledge creating means 108 include rules that are not used for guidance. Therefore, in order to reduce the number of rule searches accompanying the calculation of the number of contradictions, the contradiction occurrence number calculation means 109 extracts only the rules used for guidance from the risk knowledge candidates in FIG. By checking, the number of contradictions is calculated. As a result, the number of rule searches can be reduced, and the number of contradictions can be calculated at high speed.

  When the calculation of the number of contradictory occurrences is started (2001), first, a use rule extraction step 2002 is performed. The risk knowledge candidates shown in FIG. 5 created by the risk knowledge creation means 108 include rules that are not used for guidance. Therefore, the contradiction occurrence number calculation means 109 uses only the rules used for guidance from the risk knowledge candidates shown in FIG. Are extracted to create risk knowledge candidates for only the usage rules shown in FIG. Specifically, the risk knowledge candidate rules in FIG. 5 are first sorted in descending order of incidence. In this case, rule 512 has the highest incidence rate of 20%, and rule 512 (onset rate 20%), rule 513 (onset rate 19%), rule 514 (onset rate 18%), rule 517 (onset) (Rate 15%), rule 510 (onset rate 13%), rule 511 (onset rate 12%), rule 508 (onset rate 11%), rule 509 (onset rate 10%), rule 518 (onset rate 9%), The rules are sorted in the order of rule 515 (onset rate 8%) and rule 516 (onset rate 5%). Next, the rule with the highest incidence is extracted as a usage rule, and the rule including the same condition as the rule and the condition of the rule is deleted from the risk knowledge candidates in FIG. Also, rules with less than a certain level of support are deleted. In this case, rule 512 (onset rate 20%) is extracted as a usage rule, and rule 514 and rule 514 including the same condition as rule 512 are deleted. Further, if the support level condition is 0.1%, the rule 509 having a support level of less than 0.1% is also deleted. As described above, when the process of rearranging the incidences in descending order, the process of extracting the rule with the highest incidence, and the process of deleting the rules are repeated, risk knowledge candidates for only the usage rules in FIG. 12 are created. In this case, rules 512, 513, 517, 510, 511, 518, 515, and 516 are extracted as usage rules. The rule 508 is deleted because it includes the same conditions as the rule 510 and has a lower incidence than the rule 510.

  Next, a rule inconsistency check step 2003 is performed. Here, the contradiction occurrence number calculation means 109 performs a contradiction check on the risk knowledge candidates of only the use rules extracted in the use rule extraction step 2002, extracts the rules having the contradiction relationship, and calculates the number of contradiction occurrences. . Specifically, first, a rule including the condition of HbA1c is extracted from the risk knowledge candidates of only the usage rules in FIG. In this case, a rule 510 including a condition less than HbA1c5.6 is extracted from the risk knowledge candidates of only the usage rules in FIG. Next, when only the HbA1c condition of the extracted rule is changed to one level, the rule used for guidance is extracted from the risk knowledge candidates of only the use rule of FIG. In this case, when the HbA1c condition of rule 510 is changed from less than 5.6 to less than 5.6 and less than 5.8, the rule used for teaching is HbA1c5.6 to 5.8, and other conditions are the same as rule 510 or null extraction condition 1508 Take out with. In FIG. 13, the rules 517 and 511 are the rules extracted under the extraction condition 1508. Next, from the rules used for guidance extracted under the extraction condition 1508, rules that are inconsistent are extracted, and the number of inconsistencies is counted. In this case, the rule 511 having a lower incidence than the rule 510 becomes a rule inconsistent with the rule 510, and the contradiction occurrence number 1 is counted. If the above procedure is repeated by changing the rule including the HbA1c condition, the number of HbA1c conflicts is calculated.

  In this case, the rules in contradiction are rule 510 and rule 511, rule 512 and rule 513, rule 515 and rule 516, and the number of contradictions in HbA1c is 3. The calculated number of occurrences of contradiction is recorded in the database 102 in the format shown in FIG. In this case, the blood sugar level thresholds 100 and 110, the HbA1c threshold values 5.6 and 5.8 have 0 blood glucose level contradiction occurrences and the HbA1c contradiction occurrences 3 and therefore, as shown by 608 in FIG. 100 for blood glucose threshold A (602), 110 for blood glucose threshold B (603), 5.6 for HbA1c threshold A (604), 5.8 for HbA1c threshold B (605), 606 3 is recorded in the number of contradictions 607 of 0 and HbA1c. Further, the extracted rules having a contradiction are recorded in the database 102 in the format of FIG. 8 and managed by the contradiction occurrence information management means 121. In this case, the rules in contradiction are the rule 510 and the rule 511, the rule 512 and the rule 513, and the rule 515 and the rule 516. Therefore, as shown in 709 to 714 in FIG. However, rules 512 and 513 are recorded in 711 and 712, and rules 515 and 516 are recorded in 713 and 714, respectively. When the recording is finished, the calculation of the number of contradictions is finished (2004).

  Then, the contradiction occurrence number presence / absence determination step 2403 is performed again. Since the number of inconsistencies in HbA1c is not 0, HbA1c is selected as the threshold change item, and selection of the threshold change item is terminated (2404).

  Next, the support degree extraction step 1707 of FIG. 15 is performed. Here, first, the information acquisition unit 112 acquires the risk knowledge candidates of FIG. 5 managed by the risk knowledge candidate management unit 120. Next, the support level extraction means 126 extracts the support level of the rule of the division divided by the blood glucose level and the threshold value of HbA1c from the risk knowledge candidate of FIG. In this case, the support levels of the rules divided by the blood sugar level thresholds 100 and 110 and the HbA1c threshold values 5.6 and 5.8 are extracted. The support level of the extracted rule is recorded in the database 102 in the format of FIG. 7, and is managed by the support level information management means 125. In this case, management is performed as shown by 810 in FIG.

  In the sequence diagram of FIG. 19, the health guidance support terminal 101 makes a risk knowledge candidate acquisition request 2105, acquires the risk knowledge candidate 2107 from the database 102, and supports and supports inconsistency occurrence information such as inconsistency occurrence number information and inconsistency detail information. Degree information registration 2108 is performed.

  Next, a threshold value change category selection step 1708 is performed. Here, first, the information acquisition unit 112 acquires the contradiction detailed information of FIG. 8 managed by the contradiction occurrence information management unit 121 and the support level information of FIG. 7 managed by the support level information management unit 125. Next, the threshold value change category selection unit 113 selects, as the threshold value change category, the condition of the threshold value change item that is most frequently included in the contradictory rule conditions from the contradiction detailed information in FIG. If there are a plurality of selected threshold change categories, the threshold change item category having the lowest support level is selected from the support level information shown in FIG. As a result, a contradiction occurs, and the category whose threshold value should be changed can be selected. In this case, conditions 5.6 to 5.8 of the threshold change item HbA1c are included in the rule conditions that are in a contradictory relationship (710, 711, and 714 in FIG. 8) and are the most frequently included conditions, so HbA1c is 5.6 or more. Select a category less than 5.8 as the threshold change category. If the selected threshold value change category is a category delimited by two threshold values, the process proceeds to a change threshold value determination step 1710. If the segment is divided by one threshold value, the process proceeds to threshold value search step 1709. In this case, the selected category is a category with HbA1c of 5.6 or more and less than 5.8, and is a category delimited by two threshold values (5.6, 5.8) of the threshold value change item HbA1c.

  Next, a change threshold determination step 1710 is performed. Here, first, the information acquisition unit 112 acquires the support level information of FIG. 7 managed by the support level information management unit 125. Next, the threshold value changing means 114 calculates the number of contradictions by changing the two threshold values one by one from the support level information in the direction in which the support level of the threshold value change level increases. Then, from the contradiction occurrence number information, a threshold value that reduces the contradiction occurrence number is determined as a change threshold value. Thereby, of the two threshold values of the threshold value change category, a threshold value that reduces the number of contradictions when changed can be determined.

  This will be described in detail with reference to a detailed flowchart of the change threshold determination step 1710 in FIG. 16 and FIG. When the change threshold value determination is started (1801), first, a first threshold value change direction / width determination step 1802 is performed. Here, the threshold value changing means 114 determines the change direction and the change width of the first threshold value. The change direction is determined from the support level information so that the support level of the threshold value change category is high. In this case, since the threshold change category is HbA1c5.6 to 5.8, the value of 5.6, which is the first threshold, is set to a smaller value so that the support level of this category is increased. Explaining in Fig. 9, the threshold value 5.6 (1103) of HbA1c is set to a smaller value so that the support levels 1105, 1109, and 1112 of the section divided by the threshold value 5.6 (1103) and the threshold value 5.8 (1104) of HbA1c are higher. . Here, the change width is the minimum unit of the value of HbA1c. In this case, if the minimum unit of the value of HbA1c is 0.1, the change width is 0.1.

  Next, the threshold changing step 1712 in FIG. 16 is performed. Here, the threshold value changing unit 114 changes the threshold value using the change direction and change width determined in the first threshold value change direction / width determination step 1802. In this case, the first threshold value 5.6 of HbA1c is changed to 5.5. Next, risk knowledge candidate creation step 1703 in FIG. 16 is performed. Here, the risk knowledge creating means 108 creates a risk knowledge candidate using the threshold value 5.5 of HbA1c changed in the threshold changing step 1712. Next, the contradiction occurrence number calculation step 1706 of FIG. 16 is performed. Here, the contradiction occurrence number calculating means 109 calculates the contradiction occurrence number between the blood glucose level and HbA1c for the created risk knowledge candidate. In this case, assuming that the number of contradictions in blood glucose level is 0 and the number of contradictions in HbA1c is 1, it is recorded in the database 102 as 609 in FIG. Next, the support degree extraction step 1707 of FIG. 16 is performed. Here, the support level extraction means 126 extracts the support level of the rule of the division divided by the blood glucose level and the threshold value of HbA1c from the created risk knowledge candidate. In this case, the support levels of the rules divided by the blood glucose level thresholds 100 and 110 and the HbA1c thresholds 5.5 and 5.8 are extracted. The support level of the extracted rule is recorded in the database 102 as indicated by 811 in FIG.

  Next, a second threshold change / width determination step 1808 is performed. Here, the change direction and change width of the second threshold are determined. The change direction is determined from the support level information so that the support level of the threshold value change category is high. In this case, since the threshold change category is HbA1c5.6 to 5.8, the value of 5.8 which is the second threshold value is set to a larger value so that the support level of this category is high. Explaining in Fig. 9, the threshold value 5.8 (1104) of HbA1c is set to a larger value so that the support levels 1105, 1109, and 1112 of the section divided by the threshold value 5.6 (1103) and the threshold value 5.8 (1104) of HbA1c are increased. . The change width is the same as the change width determined in the first threshold change / width determination step 1802 in order to compare the number of contradictions when the first threshold value and the second threshold value are changed with the same change width. . Next, the threshold changing step 1712 in FIG. 16 is performed. Here, the threshold value changing unit 114 changes the threshold value using the change direction and change width determined in the second threshold value change direction / width determination step 1802. In this case, the second threshold value 5.8 of HbA1c is changed to 5.9. Next, risk knowledge candidate creation step 1703 in FIG. 16 is performed. Here, the risk knowledge creating means 108 creates a risk knowledge candidate using the threshold value 5.9 of HbA1c changed in the threshold change step 1712. Next, the contradiction occurrence number calculation step 1706 of FIG. 16 is performed. Here, the contradiction occurrence number calculating means 109 calculates the contradiction occurrence number between the blood glucose level and HbA1c for the created risk knowledge candidate. In this case, assuming that the number of contradictions in blood glucose levels is 0 and the number of contradictions in HbA1c is 2, records in the database 102 as 610 in FIG. Next, the support degree extraction step 1707 of FIG. 16 is performed. Here, the support level extraction means 126 extracts the support level of the rule of the division divided by the blood glucose level and the threshold value of HbA1c from the created risk knowledge candidate. In this case, the support levels of the rules divided by the blood glucose thresholds 100 and 110 and the HbA1c thresholds 5.6 and 5.9 are extracted. The support level of the extracted rule is recorded in the database 102 as indicated by 812 in FIG.

  Next, a contradiction occurrence number comparison step 1806 is performed. Here, the number of contradictions in HbA1c when the first threshold is changed is compared with the number of contradictions in HbA1c when the second threshold is changed. If the number of contradictions is the same value, the process returns to the first threshold value change direction / width determination step 1802 to increase the value of the change width. If the number of contradictions is different, change threshold determination step 1814 is performed. In this case, the number of HbA1c conflicts occurring when the first threshold is changed to 5.5 is 1, and the number of HbA1c conflicts occurring when the second threshold is changed to 5.9 is 2. . In the change threshold determination step 1814, the threshold changing unit 114 compares the number of contradiction occurrences when the first threshold is changed and the number of contradiction occurrences when the second threshold is changed from the contradiction occurrence number information. Is determined as the change threshold. In this case, the number of inconsistencies in HbA1c when the first threshold is changed to 5.5 is 1, and the number of inconsistencies in HbA1c is 2 when the second threshold is changed to 5.9. The first threshold value is determined as the change threshold value. When the change threshold is determined, the change threshold determination is terminated (1815).

  Next, a threshold search step 1709 is performed. Here, the threshold changing means 114 determines the change direction and change width of the change threshold determined in the change threshold determination step 1710, and changes the threshold little by little in the direction in which the support level of the threshold change classification increases. The optimum threshold value that minimizes the number of contradictions is searched. As a result, the threshold at which the number of contradiction occurrences is reduced is gradually changed from the initial threshold set by the initial threshold setting means 107, so the number of contradiction occurrences is the smallest in the vicinity of the initial threshold that is easy for the operator to teach. An optimal threshold can be searched.

  This will be described in detail with reference to a detailed flowchart of the threshold search step 1709 in FIG. When the threshold search is started (2501), first, a threshold change direction / width determination step 1711 shown in FIG. 23 is performed. Here, the threshold value changing means 114 determines the change direction and the change width of the change threshold value determined in the change threshold value determining step 1710. In this case, the change direction is a direction in which the support level of the threshold value change category is increased, and the value of the first threshold value 5.6 is set to a smaller value. Further, here, the change width is increased by the minimum unit from the change width in the change threshold determination step 1710. In this case, it increases by 0.1 to 0.2.

  Next, the threshold value changing step 1712 in FIG. 23 is performed. Here, the threshold value changing means 114 changes the threshold value with the change direction and change width determined in the threshold value change direction / width determination step 1711. In this case, the first threshold value 5.6 of HbA1c is changed to 5.4. Next, risk knowledge candidate creation step 1703 in FIG. 23 is performed. Here, the risk knowledge creation means 108 creates risk knowledge candidates using the threshold value 5.4 of HbA1c changed in the threshold change step 1712. Next, the contradiction occurrence number calculation step 1706 of FIG. 23 is performed. Here, the contradiction occurrence number calculating means 109 calculates the contradiction occurrence number between the blood glucose level and HbA1c for the created risk knowledge candidate. In this case, assuming that the number of contradictions in blood glucose levels is 0 and the number of contradictions in HbA1c is 0, it is recorded in the database 102 as 611 in FIG. Next, the support degree extraction step 1707 of FIG. 23 is performed. Here, the support level extraction means 126 extracts the support level of the rule of the division divided by the blood glucose level and the threshold value of HbA1c from the created risk knowledge candidate. In this case, the support level of the rule divided by the blood glucose level thresholds 100 and 110 and the HbA1c thresholds 5.4 and 5.8 is extracted. The support level of the extracted rule is recorded in the database 102 as indicated by 813 in FIG.

  Next, a contradiction occurrence number judgment step 1715 is performed. When the number of contradiction occurrences of the threshold change item becomes 0 by changing the threshold value, or when the number of contradiction occurrences increases, the threshold search is terminated (2502), and the process proceeds to the threshold determination step 1716. Otherwise, the process returns to the threshold value change direction / width determination step 1711. In this case, the threshold value of HbA1c is changed to 5.4 and the number of contradiction occurrences becomes 0. Therefore, the threshold value search is terminated (2502), and the process proceeds to threshold value determination step 1716.

  In the sequence diagram of FIG. 19, the health guidance support terminal 101 makes a contradiction occurrence information and support level information acquisition request 2109 such as the contradiction occurrence number information and the detailed contradiction information, and obtains the contradiction occurrence information and support level information 2110 from the database 102. Acquire risk knowledge candidates that have been acquired and changed the threshold value, and information on the occurrence of inconsistency and support level information 2111.

  Next, a threshold determination step 1716 is performed. Here, first, the information acquisition unit 112 acquires the number of occurrences of inconsistency in FIG. 6 managed by the inconsistency occurrence information management unit 121. Next, the contradiction occurrence information display means 115 displays a list of the obtained contradiction occurrence number information on the output device 104 as in the contradiction occurrence number display screen 1200 of FIG. 10, and allows the operator to determine the final threshold value. . As a result, a list of the number of contradiction occurrences from the initial threshold value to the optimum threshold value is displayed, so that the operator can grasp the tendency of the contradiction occurrence number due to the threshold setting, and the number of contradiction occurrences is small in the vicinity of the initial threshold that is easy to guide. A threshold can be selected.

  In this case, the number of contradictions 608 in FIG. 6 is displayed on the conflict count buttons 1212 and 1213, the number of contradictions 609 is displayed on the conflict count buttons 1211, the number of conflicts 610 is displayed on the conflict count buttons 1214, The number 611 of contradiction occurrences is displayed on the contradiction occurrence number button 1210. The operator first presses either the contradiction occurrence number button 1212 or 1213, and if 1213 is pressed, from among the contradiction occurrence number buttons 1210 to 1212, if the operator presses 1212, the contradiction occurrence number button 1213 to When the user selects from 1214 and finally presses a threshold value determination button 1215, the final threshold value is determined. For example, first, when the contradiction occurrence number button 1213 is pressed, then 1210 is selected and the threshold value decision button 1215 is pressed, the blood glucose level threshold is 100 (1203), 110 (1202), and the threshold value of HbA1c is 5.4 (1205) , 5.8 (1208), which is the optimal threshold value where the number of contradictions between the blood glucose level and HbA1c is 0.

  In the sequence diagram of FIG. 19, the health guidance support terminal 101 makes a contradiction occurrence number information acquisition request 2112, acquires the contradiction occurrence number information 2113 from the database 102, determines a threshold value, and performs guidance risk knowledge registration 2114. .

  Next, a threshold change category presence / absence judgment step 1717 is performed. Here, it is determined from the support level information, the contradiction occurrence number information, and the detailed contradiction information whether there is still a category where the support level is low and a contradiction occurs in the category where the threshold is not changed. If there is still a category where the degree of support is low and a contradiction occurs, the process returns to the threshold change category selection step 1708;

  Next, a threshold value determination step 1718 for all items is performed. Here, it is determined whether or not the threshold values of all the medical examination items used for instruction have been determined. If not yet determined, the process returns to the threshold change item selection step 1705. If determined, the process proceeds to a risk knowledge creation step 1719 for guidance.

  Next, a risk knowledge preparation step 1719 for guidance is performed. Here, the risk knowledge creating means 108 creates guidance risk knowledge using the determined threshold value. In this case, risk knowledge for guidance is created using threshold values 100 and 110 for blood glucose levels and threshold values 5.4 and 5.8 for HbA1c. The prepared risk knowledge for guidance is recorded in the database 102 in the format shown in FIG. 14 and managed by the risk risk management means 124 for guidance.

  Next, the flow of processing from inputting the health check result to displaying the incidence will be described with reference to the flowchart of FIG. 22, FIG. 20, and FIG. This process is a process when an instructor such as a doctor or public health nurse presents an onset rate to an instruction target person such as a health check-up examinee and gives an instruction.

  FIG. 20 is a diagram illustrating an example of a guidance screen displayed on the output device 104 by the guidance content display unit 117, and is a diagram illustrating a state before inputting the health check result of the guidance target person. FIG. 21 is a diagram illustrating an example of a guidance screen displayed on the output device 104 by the guidance content display unit 117, and is a diagram illustrating a state in which an onset rate is displayed based on a health check result of a guidance target person. On the guidance screen 2201, when the condition is selected from the medical examination items 2204 to 2209 and the prediction button 2203 is pressed, the onset rate is displayed in the onset rate display column 2202. 2210 to 2212 are buttons for selecting an age, 2213 to 2215 are buttons for selecting a BMI, 2216 to 2218 are buttons for selecting a blood glucose level, 2219 to 2221 are buttons for selecting an HbA1c, and 2222 to 2223 are buttons for selecting a drinking habit Buttons 2224 to 2225 are buttons for selecting a diabetes history of a family such as parents and siblings.

  When the processing of FIG. 22 is started (2301), a medical examination result input step 2302 is performed. In the health check result input step 2302, the health check result input means 116 inputs the health check result of the person to be instructed. First, the screen of FIG. 20 is displayed on the output device 104, and the input of the health check result of the guidance subject is waited. Then, when the operator inputs the medical examination result of the person to be instructed for the medical examination items 2204 to 2209 by operating the input device 103 with buttons and input fields and presses the prediction button 2203, the medical examination result input means 116 inputs the medical examination result. Get the condition that was set. Here, the health checkup results of the subjects were age 40 to 49 (2211), BMI 25 to 28 (2214), blood glucose level 100 to 110 (2217), HbA1c5.4 less (2219), not drinking (2222) Suppose you have entered no family history (2224).

  Next, in a rule search step 2303, the information acquisition means 112 searches the rule corresponding to the medical examination result input in the medical examination result input step 2302 from the guidance risk knowledge. First, the information acquisition unit 112 searches for a corresponding rule from among the rules shown in FIG. In this case, the rule 1503 with the highest incidence is selected as the search result among the rules having the condition part that can be formed by the combination of the conditions of the input subject.

  Next, in the result display step 2304, the instruction content display means 117 displays the result on the output device 104. At this time, as shown in FIG. 21, the ages are 40 to 49 (2211), BMI 25 to 28 (2214), blood glucose level 100 to 110 (2217), less than HbA1c5.4 (2219), do not drink ( 2222), with no family history (2224) button selected, the onset rate display column 2202 displays the onset rate 10% of the rule 1502 acquired in the rule search step 2303.

  Then, the procedure returns to the medical examination result input step 2302. If the instructing operator wants to provide guidance that HbA1c is high when the risk is high, change only HbA1c from 5.4 to 5.8 (2220) and press the prediction button 2203. Then, in the rule search step 2303, the rule 1503 is acquired from the guidance risk knowledge shown in FIG. 14, and in the result display step 2304, the onset rate 14% higher than the onset rate 10% in the case of less than HbA1c5.4 is displayed.

  In this way, the occurrence rate is predicted when the onset rate is predicted by repeating this medical examination result input step 2302, rule search step 2303, and result display step 2304, and the input condition of the same person is changed. Display the incidence based on the results of the health checkup of the target person and give guidance. If it is determined to end in the end determination step 2305, the process ends (2306).

  As described above, in the health guidance support system of the present invention, the threshold value change category selection means is based on the rule having the contradiction calculated by the contradiction occurrence number calculating means and the support degree extracted by the support degree extraction means. Then, the category whose threshold value should be changed is selected, and the threshold value changing means sets the threshold value of the above category while changing the threshold value from the initial threshold value and confirming the number of contradictions and the support level. As a result, it is possible to set a threshold value that has a high degree of support in the vicinity of the initial threshold value and that has less contradiction, and has the effect of being able to create risk knowledge that presents guidance content that is easy for a leader to teach.

  Further, in the health guidance support system of the present invention, the threshold change category selection means 113 includes the threshold change items that are most frequently included in the contradictory rule conditions from the contradiction detailed information calculated by the contradiction occurrence number calculation means 109. Is selected as the threshold change category. If there are a plurality of selected threshold change categories, the threshold change item category having the lowest support level is selected from the support level information extracted by the support level extraction means 126. This causes an inconsistency and has an effect of selecting a category whose threshold value should be changed.

  Further, in the health guidance support system of the present invention, the threshold value changing means 114 calculates the number of contradictions by changing two threshold values one by one from the support level information in a direction in which the support level of the threshold change category increases. To do. Then, from the contradiction occurrence number information, a threshold value that reduces the contradiction occurrence number is determined as a change threshold value. This has the effect of being able to determine a threshold value that reduces the number of contradictions when the threshold value is changed, out of the two threshold values of the threshold value change category.

  Further, in the health guidance support system of the present invention, the threshold changing unit 114 determines the change direction and the change width of the change threshold determined in the change threshold determination step 1710, and the support for the threshold change classification is increased. Change the direction little by little and search for the optimal threshold that minimizes the number of contradictions. As a result, the threshold at which the number of contradiction occurrences is reduced is gradually changed from the initial threshold set by the initial threshold setting means 107, so the number of contradiction occurrences is the smallest in the vicinity of the initial threshold that is easy for the operator to teach. There is an effect that the optimum threshold value can be searched.

  Furthermore, in the health guidance support system of the present invention, the threshold value change item selection unit 111 calculates the number of contradictions from the disease contribution level information calculated by the disease contribution level calculation unit 110 in the order of items having a higher contribution level to the disease. , Select the item whose contradiction occurrence number is not 0 as the threshold change item. Thereby, there is an effect that the contradiction can be reduced by changing the threshold in order from an important item having a high contribution to the disease.

  Furthermore, in the health guidance support system of the present invention, the contradiction occurrence information display means 115 displays a list of contradiction occurrence number information on the output device 104 as shown in the contradiction occurrence number display screen 1200 of FIG. A critical threshold is determined. As a result, a list of the number of contradiction occurrences from the initial threshold value to the optimum threshold value is displayed, so that the operator can grasp the tendency of the contradiction occurrence number due to the threshold setting, and the number of contradiction occurrences is small in the vicinity of the initial threshold that is easy to guide. There is an effect that a threshold value can be selected.

  Furthermore, in the health guidance support system of the present invention, in order for the contradiction occurrence number calculating means 109 to reduce the number of rule searches associated with the calculation of the contradiction occurrence number, only the rules used for guidance are selected from the risk knowledge candidates in FIG. And the number of inconsistencies is calculated by checking for inconsistencies between usage rules. As a result, the number of rule searches can be reduced, and the number of contradictions can be calculated at high speed.

  Furthermore, in the health guidance support system of the present invention, the health examination result input means 116 inputs the health examination result of the person being instructed, and the information acquisition means 112 corresponds to the health examination result inputted from the risk knowledge for guidance. The guidance content display means 117 displays the onset rate of the guidance target person. As a result, the instructing operator is able to provide effective health guidance by presenting the onset rate corresponding to the health check result of the instructed person from risk knowledge for guidance with less contradiction. .

  In the above embodiment, a method is described in which the threshold changing unit 114 calculates the number of contradiction occurrences while gradually changing the threshold value of the threshold change item, and the operator determines a threshold value with a small number of contradiction occurrences from the contradiction occurrence number information. However, for example, a plurality of threshold candidates may be created by setting the threshold value of HbA1c in increments of 0.1 and the threshold value of blood glucose level in increments of 5, and the threshold value may be determined from the number of contradiction occurrences of the threshold candidates. Thus, by creating a plurality of threshold candidates, there is an effect that the operator can determine the threshold while grasping the tendency of the number of contradictions occurring when the threshold is changed in more detail.

  In the above-described embodiment, an example in which two blood glucose threshold values and two HbA1c threshold values are determined has been described. However, the number of threshold values may be any number. The threshold change category selection means 113 selects the category where the contradiction has occurred and the support level is low in the category where the threshold value has not been changed from the contradiction detailed information and support level information, and the threshold value change unit 114 selects the contradiction occurrence number information. This can be achieved by changing the threshold value for reducing the number of contradictions from the support level information in the direction of increasing the support level. By determining a plurality of threshold values, an operator who is instructed has the effect of being able to perform more detailed instruction.

  Moreover, although the example which determines the threshold value of two items of a blood glucose level and HbA1c was demonstrated in the said Example, two or more some items may be sufficient. The threshold change item selection unit 111 selects threshold change items in descending order of the contribution to the disease from the disease contribution level information, and the threshold change category selection unit 113 determines the threshold change item from the contradiction detailed information and the support level information. The threshold change section 114 is selected, and the threshold changing means 114 changes the threshold value at which the number of contradiction occurrences decreases from the contradiction occurrence number information and support level information in a direction in which the support level increases. Since the contradiction can be reduced for other various items, it is easy for the instructing operator to provide guidance.

  In the above embodiment, an example in which the threshold values of the two items of blood glucose level and HbA1c are determined has been described, but other items may be used. For example, when an item that is frequently used for teaching is used, the number of contradictions in the item can be reduced, which has the effect of facilitating teaching.

  In the above embodiment, the initial threshold value setting unit 107 displays the initial threshold value input screen to input the initial threshold value of the medical examination items that are easy for the operator to instruct. May be recorded in the database 102, and an initial threshold value may be set from the information. By doing so, there is an effect that it is possible to save the operator from having to input the initial threshold by himself / herself.

  In the above embodiment, the example in which the threshold change width determined by the threshold changing unit 114 is set as the minimum unit of threshold change items has been described. However, the operator may set an arbitrary change width. By doing so, there is an effect that a threshold value more reflecting the operator's intention can be set.

  Moreover, although the said Example demonstrated the example on condition of age, BMI, a blood glucose level, HbA1c, drinking habits, and a family history, you may use other conditions. Various items included in the medical examination information, such as various examinations, lifestyles through interviews, judgments by doctors, etc. may be used as conditions. In the above embodiment, the incidence rate of diabetes has been described as an example of the onset rate to be displayed. However, the onset rate relating to other diseases may be used. In addition to diabetes, it can be used for all diseases related to health checkup items and lifestyle, such as hyperlipidemia, hypertension, and kidney disease. Moreover, you may make it display combining several diseases as an onset rate.

  Moreover, in the said Example, although the logistic regression model was used as a method of calculating a disease contribution degree, you may use another method. For example, other statistical models such as the Cox proportional hazard model can be used. Moreover, the operator who guides may set an original contribution degree. By doing in this way, there exists an effect which can build the risk knowledge which presents the guidance content which reflected the operator's intention more.

  In the above embodiment, the case where the incidence is used as a disease risk has been described as an example, but other indicators may be used. Indicators calculated by other methods such as health and risk, statistical indicators, and all indicators related to health and illness can be used. Moreover, although the said Example demonstrated the method of using an association rule mining as a method of creating risk knowledge, you may create rule data using another mining method.

  In the above embodiment, the disease contribution calculation step 1704 has been described as being performed before the risk knowledge candidate creation step 1703, but either may be processed first. Moreover, you may process in parallel. Processing in parallel has the effect of reducing processing time.

  In the above-described embodiment, the method of inputting the medical examination result by the medical examination result input unit 116 with the button or the like has been described. However, other methods may be used. For example, a text input field may be provided and input from a keyboard or the like, or a slide bar type input I / F may be provided to set a numerical value. Various user interfaces can be used.

  In the above-described embodiment, the method in which the operator inputs the health check result using the input device 103 has been described. However, the health check result input means 116 indicates the health check result of the examinee, and the health check information in FIG. It may be acquired from the accumulated database 102. This eliminates the need for the operator to input the health check result himself.

The figure which shows the example of 1 structure of the health guidance support system of this invention. The figure which shows an example of the medical examination information which a medical examination information management means manages. The figure which shows an example of the initial threshold value information which a threshold information management means manages. The figure which shows an example of the disease contribution level information which a disease contribution level information management means manages. The figure which shows an example of the risk knowledge candidate which a risk knowledge candidate management means manages. The figure which shows an example of the contradiction occurrence number information which a contradiction occurrence information management means manages. The figure which shows an example of the support level information which a support level information management means manages. The figure which shows an example of the contradiction detailed information which a contradiction occurrence information management means manages. The figure which shows an example of the support level of the rule of the division divided by the blood glucose level and the threshold value of HbA1c. The figure which shows an example of the contradiction occurrence number display screen which displays the list of contradiction occurrence number information. The figure which shows an example of the initial threshold value input screen which inputs the initial threshold value of a medical examination item. The figure which shows an example of the risk knowledge candidate of only a use rule for instruction | indication. The figure which shows an example of the rule in a contradiction extracted from the risk knowledge candidate of only a usage rule. The figure which shows an example of the risk knowledge for guidance which the risk knowledge management means for guidance manages. The flowchart which shows an example of the flow of the process of risk knowledge preparation for guidance from medical examination information. The flowchart which shows an example of the flow of a process which determines the threshold value to change. The flowchart which shows an example of the flow of a process which calculates the contradiction occurrence number of a medical examination item. The flowchart which shows an example of the flow of a process which selects the medical examination item which changes a threshold value. The sequence diagram which shows an example of the exchange between a health guidance support terminal and a database. It is a figure which shows the example of a guidance screen, and is a figure which shows the state before inputting the medical examination result of a guidance subject. It is a figure which shows the example of a guidance screen, and is a figure which shows the state which displayed the onset rate from the medical examination result of the guidance subject. The flowchart which shows an example of the flow of a process from a medical examination result input to an onset rate display. The flowchart which shows an example of the flow of a process which searches the threshold value from which the number of contradiction generation | occurrence | production decreases.

Explanation of symbols

101 ... health guidance support terminal, 102 ... database, 103 ... input device, 104 ... output device, 105 ... CPU, 106 ... storage device, 107 ... initial threshold setting means, 108 ... risk knowledge creation means, 109 ... calculation of the number of contradictions Means 110, disease contribution calculation means 111, threshold change item selection means 112, information acquisition means 113, threshold change category selection means 114, threshold change means 115, contradiction occurrence information display means 116, medical examination Result input means, 117 ... instruction content display means, 118 ... medical examination information management means, 119 ... threshold information management means, 120 ... risk knowledge candidate management means, 121 ... contradiction occurrence information management means, 122 ... disease contribution degree information management means , 124 ... Guidance risk knowledge management means, 125 ... Support degree information management means, 126 ... Support degree extraction means, 201 ... Individual ID, 202 ... Date of consultation, 203 ... Age, 204 ... Blood glucose level, 205 ... HbA1c, 206 ... BMI, 207 ... Diabetes determination, 301 ... Item, 302 ... Threshold 1,303 ... Threshold 2, 401 ... Item, 402 ... Contribution level, 501 ... condition part, 502 ... age condition, 503 ... BMI condition, 504 ... blood glucose level condition, 505 ... HbA1c condition, 506 ... support level, 507 ... onset rate, 508-518 ... rule, 601 ... threshold candidate ID, 602 ... blood glucose level threshold A, 603 ... blood glucose level threshold B, 604 ... HbA1c threshold A, 605 ... HbA1c threshold B, 606 ... number of blood glucose level conflict occurrences, 607 ... HbA1c conflict occurrences Number, 608 to 611 ... contradiction occurrence number information, 801 to 809 ... support degree A to I, 810 to 813 ... support degree information, 702 ... contradiction detailed ID, 709 to 714 ... rule in contradiction, 1101 ... blood glucose level Threshold A: 100, 1102 ... Blood glucose threshold B: 110, 1103 ... HbA1c threshold A: 5.6, 1104 ... HbA1c threshold B: 5.8, 1105-1113 ... Supporting degree of support A to I, 1200 ... Contradiction occurred Number display screen, 1201 to 1204 ... Threshold value of blood glucose level, 1205 to 1209 ... Threshold value of HbA1c, 1210 to 1214 ... Number of contradiction occurrence buttons, 1216 to 1220 ... Number of contradiction occurrences of blood glucose level, 1221 to 1225 ... Number of contradiction occurrences of HbA1c , 1215 ... Threshold decision button, 1301 ... Initial threshold input Force screen, 1302 to 1305 ... first threshold value input field, 1306 to 1309 ... second threshold value input field, 1310 ... decision button, 1508 ... extraction condition, 1502 to 1505 ... rule, 1701 ... start risk knowledge creation for guidance, 1702 ... initial threshold input step, 1703 ... risk knowledge candidate creation step, 1704 ... disease contribution calculation step, 1705 ... threshold change item selection step, 1707 ... support level extraction step, 1708 ... threshold change category selection step, 1709 ... threshold search step , 1710 ... change threshold determination step, 1711 ... threshold change direction / width determination step, 1712 ... threshold determination step, 1706 ... contradiction occurrence count calculation step, 1715 ... contradiction occurrence count determination step, 1716 ... threshold determination step, 1717 ... threshold change Classification presence / absence judgment step, 1718 ... threshold decision decision step for all items, 1719 ... guidance risk knowledge creation step, 1720 ... guidance risk knowledge creation completion step, 1801 ... change threshold decision First step, 1802 ... First threshold change direction / width determination step, 1808 ... Second threshold change direction / width determination step, 1806 ... Contradiction occurrence number comparison step, 1814 ... Change threshold determination step, 1815 ... Change threshold determination end step, 2001 ... Contradiction occurrence calculation start step, 2002 ... Use rule extraction step, 2003 ... Rule conflict check step, 2004 ... Contradiction occurrence end step, 2401 ... Threshold change item selection start step, 2402 ... High contribution item selection step, 2403 ... Number of contradiction occurrence presence determination step, 2404 ... Threshold change item selection end step, 2102 ... Initial threshold information registration, 2103 ... Medical examination information, initial threshold information acquisition request, 2104 ... Medical examination information, initial threshold information, 2105 ... Risk Knowledge candidate registration, 2106 ... Risk knowledge candidate acquisition request, 2107 ... Risk knowledge candidate, 2108 ... Conflict occurrence information, support level information registration, 2109 ... Conflict occurrence information, support level information acquisition required , 2110 ... contradiction occurrence information, support degree information, 2111 ... threshold change risk knowledge candidate and its contradiction occurrence information, support degree information registration, 2112 ... contradiction occurrence number information acquisition request, 2113 ... contradiction occurrence number information, 2114 ... guidance risk Knowledge registration, 2201 ... Guidance screen, 2202 ... Incidence rate display column, 2203 ... Prediction button, 2204-2209 ... Health examination item, 2210-2225 ... Condition input button, 2301 ... Instruction content display start step, 2302 ... Health examination result input Step, 2303 ... Rule search step, 2304 ... Result display step, 2305 ... End determination step, 2306 ... Instruction content display end step, 2501 ... Threshold search start step, 2502 ... Threshold search end step.

Claims (4)

A health guidance support system that presents information for disease prevention and health promotion from medical examination information,
Initial threshold setting means for setting an initial threshold for classifying the items of the medical examination;
Based on the condition of the medical examination items delimited by the set threshold, the rate of occurrence indicating the ratio of the number of affected persons to the combination of the conditions and the combination, and the value obtained by dividing the number of affected persons by the number of the population A risk knowledge creating means for calculating the degree of support indicating reliability of the occurrence rate as a rule, and creating risk knowledge;
A contradiction occurrence number calculating means for extracting a rule having a contradiction in the change in incidence with respect to a change in the condition of the medical examination item for the prepared risk knowledge, and calculating the number of combinations of the rule as the number of contradiction occurrences When,
A support level extracting means for extracting the support level of the section divided by the threshold value of the medical examination item in which the contradiction occurs from the prepared risk knowledge;
Threshold change category for selecting, as a category for changing the threshold, a category having a low level of support in the condition of the medical examination item that is most frequently included in the conditions of the rule and having the contradictory relationship and the support level. A selection means;
Health guidance characterized by comprising threshold value changing means for changing the threshold value at which the number of contradiction occurrences becomes the smallest when changed from the support level and the number of contradiction occurrences in a direction in which the support level of the selected category is increased. Support system.   2. The health guidance support system according to claim 1, wherein the contribution degree is high as an item for changing a threshold value from the contribution degree calculating means for calculating the contribution degree of the medical examination item to the disease, and the contribution degree and the number of contradictions. And a health guidance support system comprising threshold change item selection means for selecting a medical examination item in which a contradiction occurs.   The health guidance support system according to claim 1 or 2, further comprising: a contradiction occurrence information display means for displaying a list of contradiction occurrence numbers calculated by the contradiction occurrence number calculation means. The health guidance support system according to claim 1, 2, or 3, from a health examination result input means for inputting a health examination result of a person to be trained, and risk knowledge created by changing a threshold value by the threshold value changing means, A health guidance support system comprising information acquisition means for acquiring a rule corresponding to a medical examination result and guidance content display means for displaying the onset rate of the rule acquired by the information acquisition means.

Images