JP2012033155A - Medical examination data processing method, medical examination data processing device and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To use fluctuation inherent in a living body and enable correlations between inspection items to be accurately reflected so as to use for highly accurate prediction and reexamination of an inspection result.SOLUTION: A medical examination data processing method executes: level sectioning processing for sectioning inspection value distributions of respective inspection items of a predetermined record group in a medical examination database having data of a statistically processable amount about the respective inspection items in medical care into a predetermined number of levels respectively so that the ratios of the numbers of inspection values belonging to the same level become the same between the respective inspection items and the respective ratios of plural levels between the maximum level and the minimum level become the same; and boundary value calculation processing for respectively calculating boundary inspection values forming the boundaries of respective levels of each inspection item sectioned by the level sectioning processing.

Description

The present invention relates to a medical examination data processing method, a processing apparatus, and a program. Specifically, because it is information that represents the state of a living body from various viewpoints, it is possible to accurately reflect the correlation between inspection items by uniformly handling the inspection values of each inspection item having a correlation. The present invention relates to a medical examination data processing method, a processing apparatus, and a program.
Test items include, for example, BMI, systolic blood pressure, diastolic blood pressure, white blood cell count, red blood cell count, hemoglobin content, Ht (hematocrit), MCV (average red blood cell volume), MCH (average red blood cell hemoglobin content), MCHC (average red blood cell hemoglobin concentration). ), Platelets, GOT, GPT, γGTP, total cholesterol, neutral fat, HDL cholesterol, blood sugar as needed.

In conventional health guidance, occupational physicians and doctors in charge look at the health checkup results of the examinees, make decisions based on their own experience, and provide guidance on the future health status using linear regression equations from past test results. A method to predict and teach
However, the method based on the experience of the industrial physician and the attending physician has a problem that it requires a great deal of labor and cost. In addition, with the method based on the linear regression equation from past test results, although the trend can be read, it cannot be predicted taking into account both physiological changes as a group and individual characteristics. There is a problem that it can not stand the prediction of.
At present, there is no method for accurately predicting the future health state by utilizing the physiological characteristics inherent in the living body and the health state of each individual.

Japanese Patent Laid-Open No. 2006-343792 (Patent Document 1) describes a certain disease among patients who satisfy a first condition at a first time and satisfy a second condition at a second time. Create a number of rules that correlate the proportion (incidence rate) of patients who develop a according to disease and condition, and when the test value of the patient is entered, check which rule is satisfied and apply The system which displays the disease and its incidence is described.
The first condition is, for example, gender: male, age: 40s, BMI ≧ 25, blood sugar level <125, snacking: etc., and the second condition corresponding to the first condition is Invariant items such as gender and age, or item values excluding predetermined items from the first condition, for example, BMI ≧ 30, blood glucose level <140, snacking: to do, and the like.
JP 2006-343792 A

In the system of Patent Document 1, since it is necessary for the creator to cover rules so that there are no omissions in consideration of various circumstances, not only is troublesome, but also dependent on the experience of the creator. The degree is increased and the accuracy is also problematic.
In addition, as a result, a large number of rules exist, so that a large load is applied to the arithmetic processing for extracting the corresponding rule from the inputted examination value of the examinee.

The present invention considers the physiological characteristics inherent in living organisms from both the group and individual perspectives, so that accurate predictions and examination results can be reexamined, current disease names can be reviewed, and unknown disease names can be searched. As a basic technology for this purpose, we provide a method for classifying the level to which the test value should belong, focusing on fluctuations inherent in the living body, and further utilizing this level classification Thus, an object of the present invention is to provide a method for accurately estimating the present or future value of another inspection item from the inspection result of one or more inspection items.
Therefore, for example, paying attention to the fact that the inspection value of each inspection item is information obtained from a unified body called a living body, it provides a method for uniformly handling the inspection value of each inspection item, Thereby, the correlation between inspection items can be accurately reflected.
Further, even when an inspection value of a certain inspection item is missing, by using an accurate correlation between inspection items, the inspection result of an inspection item (group) different from the certain inspection item can be An inspection value of a certain inspection item can be estimated with high accuracy.
Further, the future inspection result can be predicted with high accuracy from the current inspection result, and the predicted value can be corrected with high accuracy.

The configuration of the present invention is described in the following [1] to [16].
[1] Configuration 1
The ratio of the number of test values belonging to the same level is the same among the test items in the distribution of the test values in the predetermined record group in the medical examination database having the amount of data that can be statistically processed for each test item And a level division process for dividing each of the plurality of levels between the uppermost level and the lowermost level into a predetermined number of levels so that the ratio is the same.
Boundary value calculation processing for respectively obtaining boundary inspection values that give boundaries of each level of each inspection item classified by the level classification processing;
The medical examination data processing method characterized by performing.
[2] Configuration 2
In configuration 1,
The top level and the bottom level are levels of abnormal test values, and each of a plurality of levels between the top level and the bottom level is a level of normal test values.
The medical examination data processing method characterized by the above-mentioned.

The predetermined record group only needs to be able to constitute data that can uniformly represent the characteristics of a large number of people without being biased toward a specific group. For example, among the enormous amount of medical examination data accumulated over several years, 1 It is possible to employ a record group in which people who have been examined only once (only one examinee ID is extracted) are extracted. Desirably, it is possible to employ a record group of persons who have been examined only once and who are determined to have “no disease” at the time of the consultation. Here, “has no disease” means that no disease name is described in the disease name declaration column of the predetermined questionnaire or that there is no disease. Thus, by adopting data of a large number of persons who are healthy or at least not aware of the disease, a more accurate result can be obtained.
Test items include, for example, BMI, systolic blood pressure, diastolic blood pressure, white blood cell count, red blood cell count, hemoglobin content, Ht (hematocrit), MCV (average red blood cell volume), MCH (average red blood cell hemoglobin content), MCHC (average red blood cell hemoglobin concentration). ), Platelets, GOT, GPT, γGTP, total cholesterol, neutral fat, HDL cholesterol, blood sugar as needed.
As a ratio between the number of sections and each level, for example, for each inspection item, 2.50% (level 1), 23.75% (level 2), 23.75% (level 3), 23.75% (level 4), 23.75% (level 5), and 2.50% (level 6). In this case, the highest level is level 6 and the lowest level is level 1. The sum of both the top and bottom levels (abnormal values) is 5%, and the levels 2 to 5 (normal values) between them are each divided into 23.75%. It should be noted that the present invention is not limited to such a category, and for example, the intermediate level may be classified into three categories or less or five or more categories, or the highest and lowest levels may be increased or decreased from 2.50%.

In the case of a healthy person, due to the fluctuation inherent in the living body, any value can be taken as long as it is within the normal value range, in other words, any value within the normal value range. Can be taken with the same probability. The inventors of the present application paid attention to this.
That is, when the range of normal values is equally divided as in Configuration 1 and level values are assigned to each category,
(1) In the case of a healthy person, the combination of the level values of each inspection item (for example, the level value of 18 inspection items) is random within the range of normal values (level 2 to level 5). For this reason, many healthy persons do not represent a specific pattern, and even if hierarchical clustering is performed, no concentration appears in a specific level value combination of a specific test item group. .
On the contrary,
(2) In the case of a sick person, if the level value of a specific test item group corresponding to the disease takes an abnormal value (level 1 or level 6) and hierarchical clustering is performed, the specific test item group corresponding to the disease is specified. A concentration on the combination of level values (level 1 or level 6) appears.
Therefore,
(3) In the process of transition from health to illness, a tendency to concentrate on a combination of specific level values of specific test item groups corresponding to the illness appears, and the correlation coefficient increases.
For the reasons described above, when classified as in configurations 1 and 2, the inspection value of another inspection item can be accurately estimated or predicted from the inspection value of one or more inspection items or a combination of inspection values. The predicted value can be corrected.
In this specification, for example, when the BMI is level 5, the inspection item / level is expressed as “BMI · 5”.

[3] Configuration 3
In Configuration 1 or Configuration 2,
Inspection that prescribes the first data when the inspection item / level (or a combination of a plurality of inspection items / levels) is selected as the first data selected from the inspection items / levels classified by the level classification process The probability of occurrence of second data at a certain level of an inspection item different from the item (or a plurality of inspection items) is examined for various combinations of the first data and the second data, and the occurrence probability is equal to or greater than a predetermined value. A rule extraction process for extracting a combination of both data as a derivation rule for deriving the second data from the first data;
The medical examination data processing method characterized by performing.
In the above, in “(or a combination of a plurality of examination items / levels)”, for example, “examination items / levels” such as “maximum blood pressure · 2” and “maximum blood pressure · 3” are impossible combinations. Therefore, it is excluded in advance from the object to be examined.
In the present specification, the probability that the second data will occur when the first data is assumed is appropriately referred to as “reliability”.
For example, “80%” can be used as the predetermined value for giving the reliability threshold value, but is not limited thereto, and may be increased or decreased as appropriate.

[4] Configuration 4
In configuration 3,
In the rule extraction process, when there are two or more inspection items / levels (or a combination of a plurality of inspection items / levels) for deriving the second data of the same inspection item, the inspection that gives the maximum occurrence probability among them The item / level (or a combination of a plurality of inspection items / levels) is the first data for deriving the level of the inspection item.
The medical examination data processing method characterized by the above-mentioned.
For example, in an example of a medical examination database, when the first data “Ht · 2” and “MCHC · 3” are assumed, the second data “hemoglobin amount · 2” is derived with a reliability of 100%. However, if the second data “hemochromosome · 1” is derived from another first data with a reliability of 83%, the second data with the higher reliability (“hemochromosome · 2”) is used. It is to choose.

[5] Configuration 5
In configuration 4, further
When a plurality of inspection items and levels of each inspection item are given as a correction determination set, an inspection item / level (or a plurality of inspection items / levels) corresponding to the first data of any derivation rule extracted by the rule extraction process. A first data extraction process for extracting the inspection item / level (or the combination of the plurality of inspection items / levels) as the first data of the correction determination set; ,
When the first data is extracted by the first data extraction process, the inspection item / level of the second data derived by the derivation rule corresponding to the first data is corrected to the inspection item corresponding to the correction determination set. Correction processing to return as data,
The medical examination data processing method characterized by performing.
The “correction determination set” given to the first data extraction process can be given from an input device such as a keyboard or a mouse, but is not limited to this. For example, it may be given in order from a predetermined list of examinees, or may be given via a LAN or an external network.

[6] Configuration 6
In configuration 5, further
Examination value of each examination item of each age section constituted by making each record of the record group constituting the time series data of the same examinee in the medical examination database belong to the age section to which the age at the visit of the record belongs, respectively Section level division processing for dividing the distribution using the boundary inspection value obtained by the boundary value calculation processing, respectively,
The probability of transition from the level of an arbitrary age section / inspection item to the level of the subsequent age section / inspection item from the level of any age section / inspection item determined from the level of the arbitrary age section / inspection item Transition probabilities based on the number of data that have transitioned to each level of the section / inspection item, and the transition probability as the transition destination provisional level from the level of the arbitrary age section / inspection item among them Calculation process,
When the age of the examinee and each test value of a plurality of test items are given as a predictive judgment set, the transition probability calculation is performed for each age of the examinee and each level of the test item corresponding to each test value. A temporary level acquisition process for acquiring a transition destination temporary level obtained by the process;
Run
In the first data extraction process, when each transition destination temporary level of the subsequent age section / examination item is acquired by the temporary level acquisition process, the transition destination temporary level set is processed as the correction determination set.
The medical examination data processing method characterized by the above-mentioned.
Record groups with the same time-series data for the same examinee include the same examinee's records across both adjacent age intervals, so that the aging of test values for each identical examinee To get. For example, it is a record group of examinees who have continued for 6 years and have taken a check once a year, and the age interval is set to 5 years.
[7] Configuration 7
In configuration 6,
The level classification process uses a record group divided by the same age section as the age section adopted by the section level classification process as a predetermined record group in the medical examination database,
The boundary value calculation process obtains a boundary inspection value for each level of each inspection item for each age section,
The medical examination data processing method characterized by the above-mentioned.
For example, in the boundary value calculation process and the section level classification process, the age sections are arranged as ˜19 years old, 20-24 years old, 25-29 years old, 30-34 years old,.

[8] Configuration 8
A storage device holding a medical examination database having a quantity of data that can be statistically processed for each medical examination item;
The inspection value distribution of each inspection item of the predetermined record group in the medical examination database has the same ratio of the number of inspection values belonging to the same level between the inspection items, and a plurality between the highest level and the lowest level. Level dividing means for dividing each level into a predetermined number of levels so that the ratio of each level is the same;
Boundary value calculation means for respectively obtaining boundary inspection values that give boundaries of each level of each inspection item classified by the level classification means;
Examination value of each examination item of each age section constituted by making each record of the record group constituting the time series data of the same examinee in the medical examination database belong to the age section to which the age at the visit of the record belongs, respectively Section level classifying means for classifying the distribution using the boundary inspection value obtained by the boundary value calculating means,
The probability of transition from the level of any age section / examination item classified by the section level classification means to the level of the subsequent age section / examination item follows from the level of the arbitrary age section / examination item. Transition based on the number of data transitioned to each level of the age section / inspection item, and the transition transition level as the transition destination provisional level from the level of the arbitrary age section / inspection item among them. Probability calculation means;
When the age of the examinee and each test value of a plurality of test items are given as a predictive judgment set, the transition probability calculation is performed for each age of the examinee and each level of the test item corresponding to each test value. Temporary level acquisition means for acquiring a transition destination temporary level obtained by the means;
When the inspection item / level (or a combination of a plurality of inspection items / levels) is selected as the first data selected from the inspection items / levels classified by the level classification means, the first data is defined. The probability of occurrence of second data at a certain level of an inspection item different from the inspection item (or a plurality of inspection items) is examined for various combinations of the first data and the second data, and the occurrence probability is a predetermined value. Rule extraction means for extracting the combination of both data as described above as a derivation rule for deriving the second data from the first data;
When each transition destination temporary level set of the subsequent age section / inspection item is acquired by the temporary level acquisition unit, any of the transition destination temporary level sets extracted by the rule extraction unit as a correction determination target The presence or absence of an inspection item / level (or a combination of multiple inspection items / levels) corresponding to the first data of the derivation rule is checked, and if applicable, the inspection item / level (or the plurality of inspection items / levels) First data extraction means for extracting the first combination data) as the first data of the correction determination target;
When the first data is extracted by the first data extraction means, the inspection item / level of the second data derived by the derivation rule corresponding to the first data is corrected to the inspection item corresponding to the correction determination target. Prediction correction means to return as data,
A medical examination data processing device characterized by comprising:

[9] Configuration 9
In configuration 8,
The level classification means uses a record group divided by the same age section as the age section adopted by the section level classification means as the predetermined record group in the medical examination database,
The boundary value calculation means obtains a boundary inspection value of each level of each inspection item for each age section,
The medical examination data processing device characterized by the above.
[10] Configuration 10
In configuration 9, further
Display control means for displaying the correction data returned by the correction means on a predetermined display device separately from the data before correction;
A medical examination data processing device characterized by comprising:

[11] Configuration 11
In Configuration 1 or Configuration 2,
Examination value of each examination item of each age section constituted by making each record of the record group constituting the time series data of the same examinee in the medical examination database belong to the age section to which the age at the visit of the record belongs, respectively Section level division processing for dividing the distribution using the boundary inspection value obtained by the boundary value calculation processing, respectively,
The probability of transition from the level of an arbitrary age section / inspection item to the level of the subsequent age section / inspection item from the level of any age section / inspection item determined from the level of the arbitrary age section / inspection item Transition probability calculation processing to be obtained based on the number of data that has transitioned to each level of the section / inspection item,
The medical examination data processing method characterized by performing.

[12] Configuration 12
In configuration 11,
The level classification process uses a record group divided by the same age section as the age section adopted by the section level classification process as a predetermined record group in the medical examination database,
The boundary value calculation process obtains a boundary inspection value for each level of each inspection item for each age section,
The medical examination data processing method characterized by the above-mentioned.
[13] Configuration 13
In configuration 12, further
When the testee's age and each test value of a plurality of test items are given as a prediction determination set, the transition probability calculation process for each age of the testee, each age section corresponding to each test value, and each level of the test item, respectively A transition destination prediction process that obtains the level having the highest transition probability among the values obtained by the above and returns a set of the levels as a prediction set,
The medical examination data processing method characterized by performing.

[14] Configuration 14
A storage device holding a medical examination database having a quantity of data that can be statistically processed for each medical examination item;
The inspection value distribution of each inspection item of the predetermined record group in the medical examination database has the same ratio of the number of inspection values belonging to the same level between the inspection items, and a plurality between the highest level and the lowest level. Level dividing means for dividing each level into a predetermined number of levels so that the ratio of each level is the same;
Boundary value calculation means for respectively obtaining boundary inspection values that give boundaries of each level of each inspection item classified by the level classification means;
Examination value of each examination item of each age section constituted by making each record of the record group constituting the time series data of the same examinee in the medical examination database belong to the age section to which the age at the visit of the record belongs, respectively Section level classifying means for classifying the distribution using the boundary inspection value obtained by the boundary value calculating means,
The probability of transition from the level of any age section / examination item classified by the section level classification means to the level of the subsequent age section / examination item follows from the level of the arbitrary age section / examination item. Transition probability calculation means for obtaining each based on the number of data that has transitioned to each level of the age section / examination item,
When the age of the examinee and each test value of a plurality of test items are given as a predictive judgment set, the transition probability calculation is performed for each age of the examinee and each level of the test item corresponding to each test value. A transition destination prediction means that obtains a level having the maximum transition probability among those determined by the means and returns a set of the levels as a prediction set;
A medical examination data processing device characterized by comprising:
[15] Configuration 15
In configuration 14,
The level classification means uses a record group divided by the same age section as the age section adopted by the section level classification means as the predetermined record group in the medical examination database,
The boundary value calculation means obtains a boundary inspection value of each level of each inspection item for each age section,
The medical examination data processing device characterized by the above.

[16] Configuration 16
A program for causing a computer to function as the medical examination data processing device according to any one of claims 8 to 10, 14, or 15.

In the configuration 1, the ratio of the number of test values belonging to the same level in the test value distribution of each test item in the predetermined record group in the medical examination database having the amount of data that can be statistically processed for each medical test item Level division processing that divides into a predetermined number of levels so that the ratio of each of the plurality of levels between the uppermost level and the lowest level is the same, and each inspection divided by the level division processing A medical examination data processing method characterized by executing a boundary value calculation process for obtaining a boundary inspection value that gives a boundary of each level of an item. In the configuration 2, in the configuration 1, the uppermost level and the lowermost level are levels of abnormal test values, and a plurality of levels between the uppermost level and the lowest level are levels of normal test values. This is a medical examination data processing method.
For this reason, it is necessary to accurately estimate, predict, or correct the predicted value of another inspection item from the inspection value or combination of inspection values of one or more inspection items. Techniques can be provided.
In addition, it is possible to provide a method for uniformly handling the inspection values of each inspection item, which is information obtained from a unified body of living organisms with different characteristics, and to accurately reflect the correlation between inspection items. It becomes.

Configuration 3 is different from Configuration 1 or Configuration 2 when one or more inspection items / levels are selected as the first data selected from the inspection items / levels classified by the level classification processing. The probability of occurrence of the second data at a certain level of the inspection items is examined for various combinations of the first data and the second data, and the combination of both data having the occurrence probability equal to or greater than a predetermined value is determined from the first data. This is a medical examination data processing method characterized by executing rule extraction processing for extracting two data as derivation rules. Further, in the configuration 4, in the configuration 3, in the case where there is the first data for deriving the second data of the same inspection item, the rule extraction processing uses the first data that gives the maximum occurrence probability in the first data. The medical examination data processing method is characterized in that the first data for deriving the level of the examination item is used. Further, in the configuration 5, in the configuration 4, when a plurality of inspection items and the level of each inspection item are given as a correction determination set, the first data of any derivation rule extracted by the rule extraction process Check if there is an inspection item / level (or a combination of multiple inspection items / levels) that meets the criteria, and if applicable, determine that the inspection item / level (or a combination of multiple inspection items / levels) is subject to correction A first data extraction process to be extracted as the first data of the set, and when the first data is extracted by the first data extraction process, an inspection item of the second data derived by the derivation rule corresponding to the first data A medical examination data processing method characterized by executing correction processing for returning a level as correction data of a corresponding examination item of the correction determination set.
For this reason, the inspection value of another inspection item can be accurately estimated from the inspection value of one or more inspection items or a combination of inspection values, or the estimated value can be corrected.

In the configuration 6, in the configuration 5, each record of the record group constituting the time series data of the same examinee in the medical examination database is assigned to the age section to which the age at the visit of the record belongs. Section level classification processing that classifies the inspection value distribution of each inspection item of each configured age section using the boundary inspection value obtained by the boundary value calculation processing, and any age section divided by the section level classification processing Probability of transition from the level of the test item to each level of the subsequent age section / test item, based on the number of data transitioned from the level of the arbitrary age section / test item to each level of the subsequent age section / test item, respectively The transition probability is determined as the transition destination temporary level from the level of the arbitrary age section / inspection item with the highest transition probability in When the calculation process, the age of the examinee and each test value of a plurality of test items are given as a prediction determination set, the age of the examinee, the age interval corresponding to each test value, and each level of the test item are respectively A temporary level acquisition process for acquiring the transition destination temporary level obtained by the transition probability calculation process, and the first data extraction process includes each transition destination temporary level of the age section / inspection item that follows the temporary level acquisition process Is acquired, the set of transition destination provisional levels is processed as the correction determination set. In the configuration 7, in the configuration 6, the level classification process uses a record group divided by the same age section as the age section adopted by the section level classification process as the predetermined record group in the medical examination database. The boundary value calculation processing is a medical examination data processing method characterized by obtaining boundary inspection values of each level of each inspection item for each age section.
For this reason, the inspection value of another inspection item can be accurately predicted and corrected from the inspection value of one or more inspection items or a combination of inspection values.

  Configuration 8 is a device that implements the processing of Configuration 6, and can provide such a device. The configuration 9 is a device that realizes the processing of the configuration 7, and such a device can be provided. The configuration 10 further includes a medical examination data process according to the configuration 9, further comprising display control means for displaying the correction data returned by the correction means on a predetermined display device in distinction from the data before correction. A device, and such a device can be provided.

In the configuration 11, in the configuration 1 or 2, further, each record of the record group constituting the time series data of the same examinee in the medical examination database belongs to the age section to which the age at the visit of the record belongs. The distribution of the inspection value of each inspection item of each age section configured by performing the section level classification process for classifying using the boundary inspection value obtained by the boundary value calculation process, and the arbitrary divided by the section level classification process The transition of the probability of transition from the age section / test item level to the subsequent age section / test item level to the following age section / test item level It is a medical examination data processing method characterized by executing a transition probability calculation process to be obtained based on each number. In the configuration 12, in the configuration 11, the level classification process uses a record group divided by the same age section as the age section adopted by the section level classification process as the predetermined record group in the medical examination database. The boundary value calculation processing is a medical examination data processing method characterized by obtaining boundary inspection values of each level of each inspection item for each age section. Further, in the configuration 13, in the configuration 12, when the testee's age and each test value of the plurality of test items are given as a prediction determination set, the age interval corresponding to the testee's age and each test value -Health examination data characterized by executing a transition destination prediction process for obtaining a level having the maximum transition probability among the levels obtained by the transition probability calculation process for each level of the examination item and returning a set of the levels as a prediction set It is a processing method.
For this reason, the inspection value of another inspection item can be accurately predicted from the inspection value of one or more inspection items or a combination of inspection values.

  The configuration 14 and the configuration 15 are devices that realize the processing of the configuration 13, and such devices can be provided.

  Since the configuration 16 is a program for causing a computer to function as the medical examination data processing device of any of the configurations 8 to 10, 14, or 15, the program that implements the device of the configurations 8 to 10, 14, or 15 Can be provided.

The flowchart which shows the procedure of medical examination data processing & prediction processing. The block diagram (a) which shows the structure of the apparatus of embodiment, and explanatory drawing (b) of the technique of a level division. Explanatory drawing (a) which shows the example of a structure of a medical examination database, and explanatory drawing (b) which shows the example which levels a test value. Explanatory drawing which illustrates the probability distribution according to the age range of a test value, and a transition probability about the case of male and triglyceride (before derivation rule application). Explanatory drawing (a) which shows the extraction example of a derivation rule, and explanatory drawing (b) which shows the example which replaces the level of a transition destination with the 2nd data of a derivation rule. The explanatory diagram of FIG. 4 shows after the derivation rule is applied. Explanatory drawing which shows the change of a display at the time of selecting arbitrary age sections and levels (45-49 years old and level 6) in explanatory drawing of FIG. Change the display when any age section / level (45-49 years old / level 6) is selected in the explanatory diagram of FIG. 4, part of the section division map is hidden, and the current prediction is also shown Explanatory drawing displayed and shown.

Embodiments of the present invention will be described with reference to the drawings.
The method / apparatus of the embodiment is realized in the computer system shown in FIG. That is, it is realized by the control device 10 reading a predetermined program from the ROM 13 or the storage device 11 as necessary and executing the procedure shown in FIG.

  A medical examination database having a quantity of medical examination data that can be statistically processed is stored in the storage device 11 constituted by a hard disk or SSD. Various applications such as an application for realizing the medical examination data processing function of the present application may be stored as necessary.

A configuration example of the medical examination database is shown in FIG.
In the example shown in the drawing, 18 examination items (BMI, systolic blood pressure (SBP), diastolic blood pressure (DBP), white blood cell count, red blood cell count, hemoglobin amount (Hb) are associated with a unique medical examinee ID and date of medical examination. ), Hematocrit (Ht), MCV, MCH, MCHC, platelets, GOT, GPT, γGTP, total cholesterol, neutral fat, HDL cholesterol (in the figure, abbreviated as HDL), blood glucose at any time). These inspection items are examples, and may be increased or decreased as appropriate, such as adding other inspection items. In another table (not shown), attributes of the examinee such as age and sex are recorded in association with a common examinee ID.

The control device 10 has a known configuration having a CPU and the like.
In the example of FIG. 2A, the storage device (hard disk) 11 is connected to the control device 10, but may be connected via a LAN or the like, for example. Also, the input device is not limited to the keyboard / mouse shown in the figure. For example, the configuration is such that a medical examination result recorded on an IC card is read and stored in a medical examination database, or input via a LAN (or the Internet). A configuration may be used in which the health check results to be stored in the health check database. The same applies to the input of data such as the person to be predicted (inspection value of each inspection item; original data of the correction determination set, original data of the prediction determination set). Further, the output device is not limited to the display, and may be a printer or a speaker. In short, the hardware configuration is arbitrary as long as the configuration of FIG. 1 can be executed.

A procedure for realizing the functions of the apparatus according to the embodiment will be described with reference to FIG.
First, the data of the record which has the item value of each test | inspection item shown to Fig.3 (a) from a medical examination database is acquired. Moreover, the attribute (age / gender) data of each examinee associated with the examinee ID of each record is also acquired from a table (not shown).
In this example, 1-time medical examination data (number of records: 64194 cases) and 6-time medical examination data (total number of examinee IDs: 12099) are used as the medical examination data.

One-time medical checkup data is data in which a record of the ID of the medical check-up exists only once in any one of the six years in which the data has been accumulated, and “has no disease” at the time of the medical check-up. It is the data of the determined person. Here, “has no disease” means that no disease name is described in the disease name declaration column of the predetermined questionnaire or that there is no disease.
The first visit data is classified by age section and sex in the level classification process (S21), and is classified into a predetermined number of levels for each examination item (= discretized), and further, boundary inspection at each level A value is calculated. In the claims, the case of level division for each age section is particularly described as section level division processing, and the case of level division without limiting the age section is simply described as level division processing. Each calculated boundary inspection value is used for leveling (= discretization) of the 6-time consultation data. Moreover, the single consultation data classified according to age section and sex and leveled for each examination item is provided to the rule extraction process (S41). Details of the rule extraction process (S41) will be described later.

The 6-time medical examination data is data in which one record of the medical examiner ID exists for each of the six years in which the data is accumulated. The data for the 6th visit, unlike the data for the 1st visit, does not consider the presence or absence of a disease report at the time of the visit. In addition, when there are a plurality of records having the same “year” of the consultation date with the same examiner ID, the data of the examinee ID is excluded.
The 6th consultation data is classified by age section and gender in the level classification process (claimed section level classification process; S21), and each level calculated by leveling the single consultation data in the same age section, respectively. Using the boundary inspection value, the inspection items are leveled (= discretized) for each inspection item. Further, the 6-time consultation data classified by age section and sex and leveled for each examination item is subjected to a transition probability calculation process (S31). Details of the transition probability calculation process (S31) will be described later.

  Note that the “grouping of the 1-time consultation data and the 6-time consultation data” in step S13 can be performed by checking the ID of the patient and the date of reception.

  In step S21, as described above, the first visit data is classified by age section and sex, and each classification is classified into a predetermined number of levels for each examination item. In addition, the 6 visit data is classified by age section and sex, and each classification is leveled using the corresponding age section / sex / level boundary test values obtained by leveling the 1 visit data. .

  As shown in Fig. 2 (b), the level of any age section / sex / test item in the single consultation data starts from the test value of the minimum value of the age section / sex / test item. , Inspection values included up to 2.50%, inspection values included up to 26.25%, inspection values included up to 50.00%, inspection values included up to 73.75%, inspection values included up to 97.50% The inspection values included up to 100.00% are obtained by obtaining each. 2. 50% or less is an abnormal value on the lower test value side, and more than 97.50% is an abnormal value on the higher side. Moreover, it is larger than 2.50% and 97.50% or less is the range made into a normal value. As the age section, in this example, the increment of 5 years is adopted. That is, from 5 to 19 years old, 20 to 24 years old, 25 to 29 years old, 30 to 35 years old, etc. are adopted in increments of 5 years (see FIG. 4).

As described above, in the present invention, the range of normal values (2.50% to 97.50%) is divided so that the number of inspection values included in each level is equal.
The reason will be explained.
In the case of a healthy person, due to the fluctuation inherent in the living body, any value can be taken as long as it is within the normal value range, in other words, any value within the normal value range. Can be taken with the same probability. The inventors of the present application paid attention to this.
That is, when the range of normal values is equally divided as described above and level values are assigned to each category,
(1) In the case of a healthy person, the combination of the level values of each inspection item (for example, the level value of 18 inspection items) is random within the range of normal values (level 2 to level 5). For this reason, many healthy persons do not represent a specific pattern, and even if hierarchical clustering is performed, no concentration appears in a specific level value combination of a specific test item group. .
On the contrary,
(2) In the case of a sick person, if the level value of a specific test item group corresponding to the disease takes an abnormal value (level 1 or level 6) and hierarchical clustering is performed, the specific test item group corresponding to the disease is specified. A concentration on the combination of level values (level 1 or level 6) appears.
For this reason,
(3) In the process of transition from health to illness, a tendency to concentrate on a combination of specific level values of specific test item groups corresponding to the illness appears, and the correlation coefficient increases.
Therefore, when classified as described above, the inspection value of another inspection item can be accurately estimated from the inspection value of one or more inspection items or a combination of inspection values, or the predicted value can be accurately corrected. Based on this, it can be expected that reexamination is instructed, that it is necessary to review the current disease name, and that an unknown disease name can be searched.

  Each boundary test value obtained by leveling the data of the first visit is used for leveling the same age section / sex / test item of the data for the sixth checkup. For example, the 40-44-year-old male visit data and the level of border test values obtained by leveling triglycerides in males are converted to the level of 40-44-year-old male males and triglyceride data obtained in 6 visits. Used for. In this way, the age section, sex, and test items are aligned, so even if there is a deviation in the standard values of test values depending on age or gender, a large amount of data can be obtained from 6 times of consultation data, which has a relatively small amount of data. Can be leveled with good accuracy equivalent to single-visit data. In other words, the above (1) to (3) can be reflected on the six-time consultation data in which the same person's consultation data is always included in adjacent age sections. In other words, in the process of transitioning from health to illness, it is possible to accurately use the phenomenon that a concentration tendency appears in a combination of specific level values of specific test item groups according to illness in the 6th consultation data. become. In other words, the correlation of the combination of the specific level values of the specific examination item group can be used with high accuracy in the 6th consultation data.

  Thus, the ratio of the number of test values included in the same level to the number of all test values in the age section, sex, and test items is the test value of each age section, each sex, and each test item of the single consultation data, Each inspection item is divided into a predetermined number of levels (six in this example) so that the ratio of each level within the range that is a normal value is the same so as to be the same between the inspection items, After using this to divide the 6-time consultation data, next, data is created by replacing the test values of both consultation data with the corresponding levels (discrete values). As a result, it is possible to uniformly handle each examination value of each age section, each sex, and each examination item of the 1-time consultation data and the 6th consultation data by “examination item / level”. An example in which the inspection value is replaced with the corresponding level is shown in FIG.

  Furthermore, regarding the first consultation data, the probability distribution of each examination item by age section and sex is obtained, and a section division diagram expressed by levels (discrete values) is created. The section division diagram is a diagram in which each age section is classified by level and arranged in order of age sections. FIG. 4 exemplifies a section division diagram created by obtaining a probability distribution of each level of each age section in the case of male / neutral fat. Level 1 is the low inspection value side and level 6 is the high side. In FIG. 4, although intervals are provided between adjacent age sections, they may be in close contact.

In this example, the section division diagram as shown in FIG. 4 is output to the display, and the transition probability and the transition destination are displayed on the section division diagram. That is, a transition mode diagram is displayed. For example, an arrow obtained by subtracting the magnitude of the transition probability when transitioning from a certain level in a certain age section to each level in the subsequent age section from a certain level in the certain age section to each level in the subsequent age section You may make it express with the thickness of. Alternatively, it may be represented by the magnitude of the brightness of the arrow or the line type (solid line, broken line, dotted line, etc.) of the arrow, and may be displayed in consideration of the difference in thickness. In FIG. 4, the vertical axis is a probability distribution, but the probability distribution on the vertical axis may be converted into a test value and displayed so that the actual test value can be understood. Good. That is, it is needless to say that the display may be appropriately improved so as to provide a user-friendly display.
Details of the transition destination display will be described later.

In step S31, a transition probability calculation process is executed using the 6th consultation data.
As described above, in the six-time medical examination data, there is one record having the same medical examiner ID in each of the six years in which the data is accumulated. Therefore, when an age section is set in increments of 5 years as in this example, the adjacent age sections (eg, 40 to 44 years, 45 to 49 years) always include data of the same examinee. . For this reason, if 6 times of medical examination data are used, it is possible to accurately estimate the aging of each person's test value. That is, the above-mentioned (1) to (3) are reflected in the 6th consultation data. For this reason, it is possible to accurately estimate the secular transition, and thus it is possible to accurately estimate the probability of transition from any level in any age section to each level in the subsequent age section. For example, extremely accurate knowledge can be given in terms of finding diseases early.

The transition probabilities are calculated for each test item, age section, and level for each data item belonging to a certain test item, age section, and level. This is done by counting the number of data to which it belongs.
For example, in the case of triglycerides / male, there are 213 data (not shown) of 6-time examination data belonging to 40 to 44 years old / level 4, but 41 cases (= 19.2 in the age section 45 to 49 years old). %) Is Level 6, 64 (= 30.0%) is Level 5, 51 (= 23.9%) is Level 4, 31 (= 14.6%) is Level 3, 26 (= 12) .2%) belongs to level 2, and 0 cases (= 0%) belong to level 1. As a result, estimation as shown in the figure becomes possible.
Based on the counting result, in step S31, the transition probability to each level of the subsequent age section starting from an arbitrary age section / level of each inspection item is calculated.

In step S41, a rule extraction process is executed. As the data, single visit data classified by age section and sex in the level classification process (S21) and classified into a predetermined number (six) levels for each examination item are used.
As described above, in the case of a sick person, concentration appears in a combination of specific level values of a specific examination item group corresponding to the disease. Also, a concentration tendency appears in the process of transition from health to illness. Therefore, when the transition to the disease has begun, it is possible to use the correlation between the specific level values of the specific examination item group with high accuracy.
Using this, in the rule extraction process, when a certain inspection item / level or a combination of two or more certain inspection items / levels (first data) is assumed, any of the first data based on the assumption Combination of the first data and the second data when the probability of occurrence of a certain level of the inspection item different from the inspection item (inspection item / level of the second data) is a predetermined value (eg, 80%) or more Are extracted as derivation rules for deriving the second data from the first data.

  For example, as shown in FIG. 5A, “the probability that the lowest blood pressure · 1 will occur when the highest blood pressure · 1 is assumed is 83.33% (confidence; see))”, which is a predetermined value (80 %) Or more. For this reason, a derivation rule of “maximum blood pressure · 1 as first data and minimum blood pressure · 1 as second data” is extracted. Similarly, when “Ht · 2 and MCHC · 3 are assumed, the probability of occurrence of hemoglobin amount · 2 is 100%, which is equal to or higher than a predetermined value (80%. Therefore,“ Ht · 2 and MCHC · 3 as first data and hemoglobin amount · 2 as second data ”is derived.

  By appropriately using such a derivation rule, the inspection value (level) of another inspection item having correlation is accurately estimated or corrected from the inspection value (level) of the inspection item having correlation or a combination thereof. It becomes possible. For example, it is possible to accurately correct the magnitude relationship of the transition probability to each subsequent level calculated in the transition probability calculation process (S31) (eg, which level has the highest transition probability). It becomes. In addition, it is possible to instruct reexamination that there is doubt about the examination result of the examinee (correction judgment set of claims), or to suggest that the current disease name needs to be reexamined. Furthermore, when a concentration that does not apply to a known disease name appears, it can be expected to search for an unknown disease name.

In the rule extraction process, combinations that are not possible as the first data, that is, combinations having the same inspection item and different levels (eg, Ht · 2 and Ht · 3) are considered as candidates for the first data. Are excluded in advance. As a result, it is possible to prevent the occurrence probability from being wasted and to reduce the load on the computer.
In addition, even when the probability that the first data and the second data occur at the same time (see support) is extremely small (eg, smaller than 3%), it is excluded from the calculation target of the occurrence probability and the load is reduced. .

In step S51, a prediction process is executed.
First, data of a prediction target person (predictive judgment set of claims) is acquired. As this data, each test value of the same test item (refer to Drawing 3 (a)) and a candidate for prediction (age, sex) can be mentioned as a medical examination database. Note that not all inspection values for the same inspection items as the medical examination database may be provided. That is, only some inspection values may be used. Further, the prediction target person may be a real examinee or a virtual examinee.

  The prediction target person data may be input from an input device such as a keyboard or a mouse, or may be configured to read data such as an IC card. Moreover, the structure which acquires the data input via LAN and also the internet may be sufficient. Alternatively, a configuration may be adopted in which prediction target persons are read out in a predetermined order from a list of medical examiners recorded in advance on a hard disk or the like and acquired in order.

  Next, the inspection value of each inspection item acquired in this way is leveled. That is, each is replaced with the corresponding level. As the boundary inspection value for leveling, the boundary inspection value corresponding to the age and sex of the prediction target person is used among the boundary inspection values calculated based on the data received once in step S21 (level classification processing). .

  Next, the probability of transition from the current level to each level of the subsequent age section is acquired. The current level is a level to which the test value of the prediction target person belongs. The subsequent age section is an age section to which the age of the prediction target person five years later belongs. As the transition probability to each level of the subsequent age section, each value calculated based on the 6th consultation data in step S31 (transition probability calculation processing) is acquired. Further, the transition destination level having the maximum transition probability among the transition probabilities is held on the memory 12 as a transition destination temporary level.

The transition probability to each level in the subsequent age section will be described.
For example, when the age of the prediction target is 42 years (age section: 40 to 44 years) and the level corresponding to the test value of the test item “neutral fat” is “4”, it is indicated by a black circle in FIG. The current position is the current level. Starting from this position, the probability of transition to each level in the subsequent age section (45 to 49 years old) is 19.2% for level 6, 30.3% for level 5, 23.9% for level 4 and level. 3 is 14.6%, level 2 is 12.2%, and level 1 is 0.0%. These are calculated on the basis of the count values in the transition probability calculation process in step S31 and are stored in the memory 12.

In this case, the transition probability has the maximum transition rate of 30.0% in the case of transition to level 5, so this level 5 is held as a transition destination temporary level. In FIG. 4, the arrow toward level 5 with the maximum transition probability is shown as the thickest. That is, in FIG. 4, the thickest solid line arrow, the thick solid line arrow, the thin solid line arrow, the thin broken line arrow, the thin dotted line arrow, and the thin two-dot chain line arrow are shown in descending order of the transition probability. As the starting point of the arrow, for example, the center position of the level can be set, but is not limited thereto. For example, as described above, since the probability distribution and / or the actual inspection value can be displayed on the vertical axis in FIG. 4, the height of the inspection value of the prediction target person may be set as the starting point.
The display of FIG. 4 is as described above, but in FIGS. 6 to 8 to be described later, the line type and thickness of the arrow, the position indicated by the arrow, or the like is appropriately changed according to the correction using the rule. Has been.

  When the transition destination level (transition destination temporary level) having the maximum transition probability is determined, correction based on the derivation rule is performed next. That is, the transition destination temporary level (the transition destination level having the maximum transition probability) obtained based on each inspection value of the prediction target person or the combination of the transition destination temporary levels corresponds to the first data of the derivation rule. When there is something, the second data (inspection item / level) derived using the derivation rule is corrected to replace the transition destination provisional level in which the second data and the inspection item are the same.

  For example, in the case where ““ Ht · 2 ”,“ MCHC · 3 ”, and“ hemoglobin amount · 3 ”are determined as the transition destination temporary level based on the test value of each test item of the prediction target person, FIG. As shown in a), there is a derivation rule with ““ Ht · 2 ”and“ MCHC · 3 ”as the first data and“ hemoglobin amount · 2 ”as the second data. For this reason, in this case, “hemoglobin amount · 3” held as the transition destination provisional level is replaced with the second data “hemoglobin amount · 2”. This state is shown in FIG.

Such correction will be described with reference to a section diagram of the neutral fat.
FIG. 6 shows a state in which the thickness of the arrow is changed according to the derivation rule as originally shown in FIG. That is, the maximum transition probability level (transition destination provisional level) from the level 4 of the triglyceride age section 40 to 44 years to the subsequent age section is the level 5 as shown in FIG. There is a rule in which “neutral fat · 4” is the second data, and the first data of the derivation rule is one or more transition destination temporary levels obtained based on each test value of the prediction target person Therefore, “neutral fat 5” is replaced with “neutral fat 4”, and as a result, the thickness of the arrow is changed as shown in FIG.

In the above description, the case where the display is switched from FIG. 4 to FIG. 6 has been described. However, in the case where the correction based on the derivation rule is automatically executed immediately after the transition destination temporary level is calculated. The display of FIG. 4 is not actually performed, and the display of FIG. 6 is immediately performed. Further, the switching from FIG. 4 to FIG. 6 may be configured to be performed in response to a user instruction input.
Moreover, the display of FIG. 4 and FIG. 6 shows an example. For example, the thickness and line type of the arrow in FIG. 6 are the same as those in FIG. 4, and the level changed by the derivation rule is highlighted instead. (For example, highlight display) or the like may be shown to the user.

Examples of the display mode corresponding to the transition probability and the mode of displaying the correction based on the derivation rule are as follows. That is,
As a display mode according to the transition probability, for example,
(A) A display mode in which an arrow is drawn from the calculation starting point to the transition destination level, and the thickness is increased as the transition probability increases.
(B) A display mode in which the line type of the arrow is represented by a solid line, a broken line, a dotted line, a chain line, or the like according to the transition probability.
(C) A display mode in which the brightness and color of the arrows are more conspicuous as the transition probability increases.
(D) A display mode in which the display of the level area of the transition destination is changed according to the transition probability. For example, a display mode in which, for example, a highlight display is made so that the level of the transition probability is higher.
(E) A display mode in which numerical values of transition probabilities are displayed alongside. Further, in this case, a display mode in which typefaces and points are changed according to the magnitude of the transition probability.
Can be illustrated.

Moreover, as a predetermined aspect incorporating the correction, for example,
(A) Both the transition destination provisional level (the level with the highest transition probability) and the level (modified level) indicated by the second data of the derivation rule are clearly distinguished and displayed so that the meaning of both can be understood. A mode of (highlighted).
(B) A mode in which the highlighting of the transition destination temporary level (the level with the highest transition probability) is stopped, and the level (modified level) indicated by the second data of the derivation rule is highlighted as a level having a high probability of transition.
(C) In the above (B), the transition probability is high but the transition probability is different from the maximum level (transition destination temporary level).
(D) In (B) above, not only the highlighting of the transition destination temporary level (the level with the highest transition probability) is stopped, but also the display based on the transition probability to another level is stopped, and the second derivation rule A mode in which only the level indicated by the data (the level after correction) is highlighted as a level having a high probability of transition.
(E) If the transition destination temporary level (the level with the highest transition probability) and the level (modified level) indicated by the second data of the derivation rule are adjacent to each other, the boundary portion is set as the destination of the arrow. To display.
Etc. can be illustrated.
Note that a display mode other than the above may be used as the display mode corresponding to the transition probability and the predetermined mode incorporating the correction.

In the description of the correction based on the derivation rule, the corrected “inspection item / level” may constitute the first data of another derivation rule or a part thereof. In such a case, the transitional provisional level of another inspection item is replaced by the second data derived by the derivation rule, and the fluctuation has spread one after another. It is also assumed that the “test item / level (the amount of hemoglobin in FIG. 5 (b) / 2)” is replaced again. In order to prevent such an indefinite state due to the circulation phenomenon, this device is configured so that the “inspection item / level” replaced by the second data of the derivation rule is not used as the first data of the derivation rule. ing.
Instead of this, for example, a configuration such as “stop applying the derivation rule when replacement is attempted again” may be employed.

  Further, in this apparatus, when any “inspection item / level” is selected, the system is configured to process the “inspection item / level” as a component of the first data for the inspection item. It is composed. In other words, the system is configured to process the inspection item by regarding the “inspection item / level” as a transition destination temporary level.

This will be described with reference to FIG. 4 (example of neutral fat).
In FIG. 4, the transition destination temporary level is “level 5”. At this time, for example, as shown in the figure, when the user selects “level 6”, the “level 6” is used for the test item “neutral fat” and each transition is made for other test items. Derivation rules are applied using the preliminary level. That is, it is checked whether or not there is first data consisting of one or two or more “examination items / levels” including “neutral fat · 6” alone or “neutral fat · 6”. The level of the same inspection item as the second data is replaced with the second data corresponding to the first data. In addition, when the indeterminate state due to circulation, that is, the “inspection item / level” modified as corresponding to the second data constitutes the first data of another derivation rule or a part thereof, further derivation In order to prevent a situation in which the originally selected “neutral fat · 6” is corrected due to the application and correction of the rule and its spread, the apparatus selects the “inspection item / level” selected by the user. ”(In this case,“ Neutral fat 6 ”). Instead of such control, control as described above, that is, control for fixing “inspection item / level” corrected as corresponding to the second data may be employed.

In the above-described user selection (here, “neutral fat · 6” selection), the transition probability to the subsequent age range starting from the “test item / level” selected by the user is also displayed.
This will be described with reference to FIGS.
In FIG. 4, the transition destination provisional level is “neutral fat · 5”. At this time, for example, when the user selects “neutral fat · 6”, the transition probability to each level of the subsequent age section is acquired from the “neutral fat · 6” as a starting point. As shown, the transition to each level is displayed with arrows of thickness and line type according to the acquired transition probability.

  In this way, the user appropriately selects the “inspection item / level” and, for example, in the current situation, the probability of transition to “level 5” is the highest five years ahead, but overdue etc. As a result, for example, if it is assumed that the level has transitioned to “level 6” five years ahead, what other test items have changed to “level 6” at the point five years ahead In addition to grasping the influence, it is also possible to grasp in an easy-to-understand manner how the transition will occur thereafter.

FIG. 8 shows an example in which FIG. 7 is displayed differently.
In FIG. 8, a part of the section division diagram is not displayed.
Further, in FIG. 8, the transition to the selected “level 6” is indicated by a two-dot chain thick arrow as an assumed prediction (maximum possibility), and the transition destination from the “level 6” is also indicated by a two-dot chain thick line. It is indicated by an arrow (maximum possibility) and a two-dot chain line arrow (next possibility).
Further, in FIG. 8, the transition probability “level 5” was the maximum from the inspection result of each inspection item of the prediction target person, but “level 4” was derived in the modification by the derivation rule, so the boundary between both levels Is set as the arrow destination.

Such an output can be output not only by a display but also by a printer, for example. Further, it may be configured to transmit to the inquiry source terminal device via the LAN or the Internet.
As a result of calculating the transition probability with the apparatus of the present application and applying the result corrected by the derivation rule to the data of a large number of medical examiners, it was confirmed that good accuracy was obtained.

  In the above description, the example in which the derivation rule is applied to the prediction destination (transition destination temporary level set) has been described. However, the derivation rule can also be applied to the current inspection result (correction determination set in claims). Moreover, as an advanced type, it is possible to instruct the review of the test result or to suggest that the review of the current disease name is necessary. Furthermore, it can be expected to be used for searching for an unknown disease name or the like when an unusual concentration appears.

10 Control device (equipped with CPU)
11 Storage devices (hard disks, etc.)
12 Storage device (RAM)
13 Storage device (ROM)
16 Input devices (keyboard, mouse, etc.)
17 Output device (display, etc.)
18 LAN

Claims (16)

The ratio of the number of test values belonging to the same level is the same among the test items in the distribution of the test values in the predetermined record group in the medical examination database having the amount of data that can be statistically processed for each test item And a level division process for dividing each of the plurality of levels between the uppermost level and the lowermost level into a predetermined number of levels so that the ratio is the same.
Boundary value calculation processing for respectively obtaining boundary inspection values that give boundaries of each level of each inspection item classified by the level classification processing;
The medical examination data processing method characterized by performing. In claim 1,
The top level and the bottom level are levels of abnormal test values, and each of a plurality of levels between the top level and the bottom level is a level of normal test values.
The medical examination data processing method characterized by the above-mentioned. In claim 1 or claim 2, further,
Inspection that prescribes the first data when the inspection item / level (or a combination of a plurality of inspection items / levels) is selected as the first data selected from the inspection items / levels classified by the level classification process The probability of occurrence of second data at a certain level of an inspection item different from the item (or a plurality of inspection items) is examined for various combinations of the first data and the second data, and the occurrence probability is equal to or greater than a predetermined value. A rule extraction process for extracting a combination of both data as a derivation rule for deriving the second data from the first data;
The medical examination data processing method characterized by performing. In claim 3,
In the rule extraction process, when there are two or more inspection items / levels (or a combination of a plurality of inspection items / levels) for deriving the second data of the same inspection item, the inspection that gives the maximum occurrence probability among them The item / level (or a combination of a plurality of inspection items / levels) is the first data for deriving the level of the inspection item.
The medical examination data processing method characterized by the above-mentioned. In claim 4, further:
When a plurality of inspection items and levels of each inspection item are given as a correction determination set, an inspection item / level (or a plurality of inspection items / levels) corresponding to the first data of any derivation rule extracted by the rule extraction process. A first data extraction process for extracting the inspection item / level (or the combination of the plurality of inspection items / levels) as the first data of the correction determination set; ,
When the first data is extracted by the first data extraction process, the inspection item / level of the second data derived by the derivation rule corresponding to the first data is corrected to the inspection item corresponding to the correction determination set. Correction processing to return as data,
The medical examination data processing method characterized by performing. In claim 5, further,
Examination value of each examination item of each age section constituted by making each record of the record group constituting the time series data of the same examinee in the medical examination database belong to the age section to which the age at the visit of the record belongs, respectively Section level division processing for dividing the distribution using the boundary inspection value obtained by the boundary value calculation processing, respectively,
The probability of transition from the level of an arbitrary age section / inspection item to the level of the subsequent age section / inspection item from the level of any age section / inspection item determined from the level of the arbitrary age section / inspection item Transition probabilities based on the number of data that have transitioned to each level of the section / inspection item, and the transition probability as the transition destination provisional level from the level of the arbitrary age section / inspection item among them Calculation process,
When the age of the examinee and each test value of a plurality of test items are given as a predictive judgment set, the transition probability calculation is performed for each age of the examinee and each level of the test item corresponding to each test value. A temporary level acquisition process for acquiring a transition destination temporary level obtained by the process;
Run
In the first data extraction process, when each transition destination temporary level of the subsequent age section / examination item is acquired by the temporary level acquisition process, the transition destination temporary level set is processed as the correction determination set.
The medical examination data processing method characterized by this. In claim 6,
The level classification process uses a record group divided by the same age section as the age section adopted by the section level classification process as a predetermined record group in the medical examination database,
The boundary value calculation process obtains a boundary inspection value for each level of each inspection item for each age section,
The medical examination data processing method characterized by the above-mentioned. A storage device holding a medical examination database having a quantity of data that can be statistically processed for each medical examination item;
The inspection value distribution of each inspection item of the predetermined record group in the medical examination database has the same ratio of the number of inspection values belonging to the same level between the inspection items, and a plurality between the highest level and the lowest level. Level dividing means for dividing each level into a predetermined number of levels so that the ratio of each level is the same;
Boundary value calculation means for respectively obtaining boundary inspection values that give boundaries of each level of each inspection item classified by the level classification means;
Examination value of each examination item of each age section constituted by making each record of the record group constituting the time series data of the same examinee in the medical examination database belong to the age section to which the age at the visit of the record belongs, respectively Section level classifying means for classifying the distribution using the boundary inspection value obtained by the boundary value calculating means,
The probability of transition from the level of any age section / examination item classified by the section level classification means to the level of the subsequent age section / examination item follows from the level of the arbitrary age section / examination item. Transition based on the number of data transitioned to each level of the age section / inspection item, and the transition transition level as the transition destination provisional level from the level of the arbitrary age section / inspection item among them. Probability calculation means;
When the age of the examinee and each test value of a plurality of test items are given as a predictive judgment set, the transition probability calculation is performed for each age of the examinee and each level of the test item corresponding to each test value. Temporary level acquisition means for acquiring a transition destination temporary level obtained by the means;
When the inspection item / level (or a combination of a plurality of inspection items / levels) is selected as the first data selected from the inspection items / levels classified by the level classification means, the first data is defined. The probability of occurrence of second data at a certain level of an inspection item different from the inspection item (or a plurality of inspection items) is examined for various combinations of the first data and the second data, and the occurrence probability is a predetermined value. Rule extraction means for extracting the combination of both data as described above as a derivation rule for deriving the second data from the first data;
When each transition destination temporary level set of the subsequent age section / inspection item is acquired by the temporary level acquisition unit, any of the transition destination temporary level sets extracted by the rule extraction unit as a correction determination target The presence or absence of an inspection item / level (or a combination of multiple inspection items / levels) corresponding to the first data of the derivation rule is checked, and if applicable, the inspection item / level (or the plurality of inspection items / levels) First data extraction means for extracting the first combination data) as the first data of the correction determination target;
When the first data is extracted by the first data extraction means, the inspection item / level of the second data derived by the derivation rule corresponding to the first data is corrected to the inspection item corresponding to the correction determination target. Prediction correction means to return as data,
A medical examination data processing device characterized by comprising: In claim 8,
The level classification means uses a record group divided by the same age section as the age section adopted by the section level classification means as the predetermined record group in the medical examination database,
The boundary value calculation means obtains a boundary inspection value of each level of each inspection item for each age section,
The medical examination data processing device characterized by the above. The claim 9, further comprising:
Display control means for displaying the correction data returned by the correction means on a predetermined display device separately from the data before correction;
A medical examination data processing device characterized by comprising: In claim 1 or claim 2, further,
Examination value of each examination item of each age section constituted by making each record of the record group constituting the time series data of the same examinee in the medical examination database belong to the age section to which the age at the visit of the record belongs, respectively Section level division processing for dividing the distribution using the boundary inspection value obtained by the boundary value calculation processing, respectively,
The probability of transition from the level of an arbitrary age section / inspection item to the level of the subsequent age section / inspection item from the level of any age section / inspection item determined from the level of the arbitrary age section / inspection item Transition probability calculation processing to be obtained based on the number of data that has transitioned to each level of the section / inspection item,
The medical examination data processing method characterized by performing. In claim 11,
The level classification process uses a record group divided by the same age section as the age section adopted by the section level classification process as a predetermined record group in the medical examination database,
The boundary value calculation process obtains a boundary inspection value for each level of each inspection item for each age section,
The medical examination data processing method characterized by the above-mentioned. The claim 12, further comprising:
When the testee's age and each test value of a plurality of test items are given as a prediction determination set, the transition probability calculation process for each age of the testee, each age section corresponding to each test value, and each level of the test item, respectively A transition destination prediction process that obtains the level having the highest transition probability among the values obtained by the above and returns a set of the levels as a prediction set
The medical examination data processing method characterized by performing. A storage device holding a medical examination database having a quantity of data that can be statistically processed for each medical examination item;
The inspection value distribution of each inspection item of the predetermined record group in the medical examination database has the same ratio of the number of inspection values belonging to the same level between the inspection items, and a plurality between the highest level and the lowest level. Level dividing means for dividing each level into a predetermined number of levels so that the ratio of each level is the same;
Boundary value calculation means for respectively obtaining boundary inspection values that give boundaries of each level of each inspection item classified by the level classification means;
Examination value of each examination item of each age section constituted by making each record of the record group constituting the time series data of the same examinee in the medical examination database belong to the age section to which the age at the visit of the record belongs, respectively Section level classifying means for classifying the distribution using the boundary inspection value obtained by the boundary value calculating means,
The probability of transition from the level of any age section / examination item classified by the section level classification means to the level of the subsequent age section / examination item follows from the level of the arbitrary age section / examination item. Transition probability calculation means for obtaining each based on the number of data that has transitioned to each level of the age section / examination item,
When the age of the examinee and each test value of a plurality of test items are given as a predictive judgment set, the transition probability calculation is performed for each age of the examinee and each level of the test item corresponding to each test value. A transition destination prediction means that obtains a level having the maximum transition probability among those determined by the means and returns a set of the levels as a prediction set;
A medical examination data processing device characterized by comprising: In claim 14,
The level classification means uses a record group divided by the same age section as the age section adopted by the section level classification means as the predetermined record group in the medical examination database,
The boundary value calculation means obtains a boundary inspection value of each level of each inspection item for each age section,
The medical examination data processing device characterized by the above.   A program for causing a computer to function as the medical examination data processing device according to any one of claims 8 to 10, 14, or 15.

Images