JP2017117469A - Risk evaluation method, risk evaluation device, and risk evaluation program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide information which becomes a reference for determining the priority of dealing by evaluating the size of a risk for an object person at the point of processing time.SOLUTION: This risk evaluation method causes a computer to execute: a step for acquiring health check-up information indicating the measured values of health check-up or check-up history during the prescribed period of a plurality of object persons and receipt information including medical costs during the prescribed period of the object persons; a step for generating a model indicating a probability that medical costs will be equal to or higher than a threshold with prescribed disease names in the receipt information by using values based on the health check-up information as explanatory variables; and a step for reading new health check-up information about the object persons and calculating risk evaluation values by using the values based on the health check-up information and the model.SELECTED DRAWING: Figure 1

Description

  The present technology relates to a risk evaluation method, a risk evaluation apparatus, and a risk evaluation program.

  Medical insurers such as health insurance associations carry out specific health checkups and specific health guidance for the prevention and early detection of lifestyle-related diseases. In order to reduce medical expenses, it is important to promote improvement of lifestyle habits by providing health guidance to insured people, but the resources of the personnel who provide health guidance are limited. Therefore, it is important to appropriately select the target persons for health guidance.

  Conventionally, a system has been proposed that automatically and efficiently creates a prediction model that can easily grasp the causal relationship with pathological changes and pathological changes using receipt information, medical examination information, etc. (For example, Patent Document 1). Specifically, the system includes a node corresponding to a random variable representing an item of shaping information based on shaping information obtained by collecting the analysis target person's receipt information and medical examination information for each analysis target person and every predetermined period. A graphical model is defined in which the probabilistic dependence between them is defined by a directed edge or an undirected edge, and a disease onset probability and medical expenses are predicted based on the graphical model. Furthermore, the dependence between nodes is calculated | required based on the random variable of each node in a graphical model, or the presence or absence of the edge between each node. Then, the system calculates the degree of influence of the item corresponding to each node in the graphical model on the medical cost based on at least one of the medical cost applied for prescription and the predicted medical cost, and the degree of influence Based on the above, the influence of medical expenses acting on each node is calculated. Further, the system calculates an inter-item force defined by repulsive force and attractive force acting between the nodes based on the dependency between the nodes. Subsequently, the system determines the coordinate arrangement of each node in the graphical model so that the energy defined by the sum of the medical cost influence force and the inter-item force decreases, and based on the coordinate arrangement of each node, Visualization of a graphical model.

  In addition, for example, in order to perform early prediction of complications in a patient in an intensive care unit, a technique has been proposed in which a mathematical expression for determining a morbidity rate is determined using a logistic function (for example, Patent Document 2). . This is an early prediction of patient complications in the ICU by detecting and using subtle patterns related to fluctuations occurring over a period of time, but using multiple continuous time series data It is assumed that.

  In addition, the guidance effect is calculated in time series for each group obtained by combining item conditions from medical examination information, guidance information, and receipt information, and an approximate straight line is calculated as a feature value from the time series guidance effect. There has also been proposed a health business support system that creates a health guidance model related to guidance priority using item conditions selected based on the amount (for example, Patent Document 3). The purpose of this is to improve the effect by evaluating the health guidance planned and implemented based on the created model and reviewing the plan.

International Publication No. 2015/068812 Special table 2013-536971 gazette JP 2007-257565 A

  Thus, although the method of calculating | requiring the future onset probability about a certain disease and the influence degree to medical expenses was proposed, the effect was not fully acquired. The major problem is that conventional methods such as analysis of causality and creation of a prediction model do not have sufficient accuracy. In addition, even if it is possible to calculate the probability of onset and the impact on medical expenses, the response method based on the results is not well established, and the environment that can be implemented is not established. In other words, medical treatment of treating after the onset remains the main response method, and although there are activities to improve life and health status through health guidance, sufficient effects are obtained. There is a current situation that is difficult to say.

  On the other hand, there is an idea of wanting to cope before getting sick without depending on medical technology, for example, active treatment such as drug administration and surgery. In other words, it is not intended to extend the life span of the life span by treatment, but to extend the healthy life span by preventive medicine including health guidance before onset. This is one of the ideas that has been advocated in recent years as one of the approaches to lead a more fulfilling life that is more human.

  However, the required resources such as cost and manpower for achieving this are limited. Therefore, in order to provide health guidance more efficiently, it is useful to extract health guidance subjects with high priorities based on the current level of risk based on health checkup results and medical expenses. The present inventors have found out.

  Therefore, the present invention has an object to grasp the health condition of the subject at the time of processing, evaluate the magnitude of risk, and present information serving as a reference for determining the priority of correspondence. .

  The risk evaluation method according to the present invention obtains, from a storage unit, medical checkup information including medical checkup values for a plurality of subjects in a predetermined period and receipt information including medical expenses for a plurality of subjects in a predetermined period. A step, a step of generating a model representing a probability that a medical cost is equal to or greater than a predetermined threshold in the receipt information, with a value based on the medical examination information as an explanatory variable, and new medical examination information of the subject Is read from the storage unit, and the computer executes a step of calculating a risk evaluation value using a value and a model based on the medical examination information.

  In this way, a model is generated based on the information related to the medical examination and the receipt information, and the risk evaluation value can be calculated based on the new medical examination information. Therefore, it becomes possible to evaluate the magnitude of the risk of the target person at the time of the processing and present information serving as a reference for determining the priority of correspondence. The medical examination information is not limited to the results of so-called regular medical examinations, but may be medical examination information received for specific items at a pharmacy or the like.

  Further, the value based on the medical examination information may be a value obtained by converting one or more measurement values in the medical examination into a score represented by a predetermined order scale, or using the obtained measurement values as they are. Also good. Specifically, such a value can be used as an explanatory variable. Further, the value based on the medical examination information and the receipt information may be a measurement value of a predetermined medical examination item in the medical examination information and the number of medical examinations with a predetermined injury / illness name in the reception information. In this way, there is no need to provide a reference for scoring, and this can be carried out more easily. Moreover, since the value is used as it is without being converted, the calculation result is in line with the actual situation.

In addition, values based on health checkup information include measured values related to hypertension, measured values related to hyperlipidemia, measured values related to diabetes, or visit history related to diseases such as a predetermined brain, heart, or circulatory system. At least one of them may be included. Furthermore, for example, the fact of intervention by a public health nurse or doctor, such as when receiving health guidance, may be digitized and used as data. Specifically, such a value may be used as an explanatory variable. These items in the medical examination information are values found by the inventor to be useful for extracting subjects whose medical expenses have exceeded a predetermined amount due to diseases such as a predetermined brain, heart, or circulatory organ. .

  As long as the object of the present invention is a method that can achieve an appropriate evaluation of the current risk level of the subject, the risk evaluation calculation method is not particularly limited, and is a linear function. Or a non-linear function. For example, when the calculation is performed using a nonlinear function, the model may be represented by a logistic function for obtaining a probability that a medical cost is equal to or higher than a predetermined threshold value with a predetermined disease name.

  Further, a risk evaluation value may be calculated for a plurality of target persons, and a step of extracting a predetermined upper number of target persons based on the magnitude of the calculated risk evaluation value may be further executed. In this way, it becomes possible to extract subjects with high priority based on the magnitude of risk.

  In addition, when the medical examination information and the receipt information in the latest period are registered in the storage unit, a step of generating a model using the medical examination information and the receipt information in the latest period may be further executed. Good. If it does in this way, a model reflecting the actual condition and the latest tendency can be generated from the medical examination information and the receipt information of the subject person, and the risk evaluation can be performed.

  The contents described in the means for solving the problems can be combined as much as possible without departing from the problems and technical ideas of the present invention. The contents of the means for solving the problems can be provided as a device such as a computer or a system including a plurality of devices, a method executed by the computer, or a program executed by the computer. The program can be executed on a network. A recording medium that holds the program may be provided.

  According to the present invention, it becomes possible to grasp the health condition of a subject at the time of processing, evaluate the magnitude of risk, and present information serving as a reference for determining the priority of correspondence. This makes it possible to efficiently provide health guidance with limited resources, and as a result, not only can it contribute to the effect of reducing medical expenses, but also for people who want to extend their healthy life span and lead a fulfilling life every day. It is also possible to meet requests.

It is a functional block diagram of a risk analyzer. It is an apparatus block diagram which shows an example of a computer. It is a processing flowchart which shows an example of a risk analysis process. It is a processing flow figure showing an example of model generation processing. It is a figure which shows an example of the conditions used for conversion to a score. It is a processing flowchart which shows an example of a risk evaluation process. It is the figure which highlighted and displayed the subject whose medical expenses are more than a predetermined threshold while sorting the subject concerning Example 1 by the onset probability P. It is the figure which highlighted and displayed the subject whose medical expenses are more than a predetermined threshold while sorting the subject concerning Example 2 by the onset probability P. It is the figure arranged in order with the highest total value based on the value of HbA1c. It is the figure which gave the score based on the value of blood glucose, blood pressure, and a lipid, and arranged in order with a high total score. It is a figure which shows the ROC curve which evaluates the onset probability of a cerebral cardiovascular disease based on this system using the test item which does not make a score, and a consultation history. It is a figure which shows the ROC curve which employ | adopts the test | inspection item of LDL cholesterol and evaluates the onset probability of cerebrocardiovascular disease.

  Hereinafter, with respect to the embodiment of the present invention, the results of analyzing the data of subjects belonging to a specific health insurance association using a logistic function will be described with reference to the drawings, but the following embodiment is an example, The present invention is not limited to the following configuration.

<Functional configuration of the device>
FIG. 1 is a functional block diagram of the risk analysis apparatus according to the embodiment. The risk analysis apparatus 1 according to the present embodiment includes a medical examination information storage unit 11, a receipt storage unit 12, a model generation unit 13, a model storage unit 14, a risk evaluation unit 15, and a result storage unit 16. Prepare.

The medical examination information storage unit 11 stores, for example, data (also referred to as “medical examination information”) indicating a result of a medical examination of subjects belonging to a predetermined group. Here, the term “predetermined” means that the information of the content that specifies the range in advance is provided in carrying out the present invention. The types of information may include ID, time, period, name of injury, medical expenses, test item, measurement value, probability of onset, process, receipt information, history of consultation, number of subjects, year, etc. As long as the object can be achieved and the invention can be carried out, it can be used, and the contents and scope thereof are not uniformly limited, and those skilled in the art can appropriately set the contents of the information. Can be implemented. In addition, the group is a group that provides medical insurance. For example, a group such as a specific health insurance association, a national health insurance association, or a mutual aid association can be appropriately set. Further, the target person belonging to the group may be all persons constituting the group or a part thereof. It is assumed that the medical examination information holds the latest fiscal year and the past fiscal year for the target person. In addition, the medical examination information includes the date of medical examination (year of medical examination) and measurement values of predetermined examination items in association with the identification information of the subject. Test items are items that can be measured as general health checkups and items that can be measured as options, such as physical examination (height, weight, etc.), visual acuity, hearing, blood pressure (systole and diastole), urine in general (protein, Sugar, occult blood, etc.), stool, blood in general (white blood cell count, red blood cell count, hemoglobin content, hematocrit, MCV, MCH, MCHC, platelet count, etc.), liver function (total protein, A / G ratio, albumin, total bilirubin, TTT) , ZTT, AST (GOT), ALT (GPT), ALP, LDH, ν-GTP, cholinesterase, etc.), pancreas (serum amylase), lipid (total cholesterol, HDL cholesterol, LDL cholesterol, neutral fat (TG), artery Sclerosis index, etc.), gout (uric acid), renal function (inulin, cystatin C, urea nitrogen, creatinine, eGFR, etc.), diabetes (glucose) Fructose, blood glucose, HbA1c (hemoglobin A1c), etc.), serum (HBs antigen, HCV-III antibody, TPHA method, RPR method, etc.), predetermined brain, heart or cardiovascular disease, blood type, blood image, electrocardiogram, It shall include at least part of chest X-ray, upper gastrointestinal tract, metabolic assessment, health guidance level, abdominal ultrasound, medical interview, comprehensive assessment, and overall assessment. In recent years, the manner of examination has diversified beyond health checkups. In addition to the above-mentioned health checkups, the health checkup information will be updated as appropriate for some test items, reflecting the results of any blood test or saliva test performed by the subject in a medical facility or pharmacy. It may be. In this case, the examination result received arbitrarily may be used in place of the medical examination information, or the examination result optionally given in addition to the medical examination information may be used. In addition to the health checkup information, if you use the results of the test you have received at your discretion, it will be easier to determine the priority of the response because you can understand the current state of health and risk assessment based on more accurate calculations. preferable. In addition, if the dependent of the health insurance association insured is the target person, the low coverage rate of the medical checkup is considered to be a problem for the dependent of the insured, and it may be difficult to use sufficient health checkup information. is assumed. Insufficient medical examination information may lead to inadequate health assessment and risk assessment accuracy. However, using the results of examinations that have been voluntarily examined, This is preferable because risk evaluation can be performed with higher accuracy.

  The receipt storage unit 12 stores information on a receipt, which is a specification of medical expenses (medical fee and dispensing fee) issued by a medical institution. The receipt information includes the name of the main injury and the medical cost in association with the identification information of the examinee. The identification information includes, for example, personal information such as name, gender, date of birth, health insurance participation information, and the like. Therefore, it is possible to add up the number of times of consultation (namely, the number of times of hospital visits) for each subject with a predetermined injury / illness name. It is assumed that the identification information of the subject person of the medical examination information and the identification information of the examinee of the receipt information are linked directly or indirectly.

  The model generation unit 13 uses the medical examination information for the past year in the medical examination information storage unit 11 and the receipt information for a predetermined period (for example, the past year) in the reception storage unit 12 to obtain a predetermined onset probability. A function (also called “model”) for calculation is generated. The predetermined onset probability is machine-learned as having developed when a medical treatment is performed for a predetermined injury and illness and the medical cost is equal to or higher than a predetermined threshold. At this time, if it is a person skilled in the art, according to the cost for health guidance or the number of subjects to be extracted, necessary information is appropriately set from the information included in the medical examination information and the receipt information, and the onset probability is determined. It can be used for calculation. The model generation unit 13 determines a parameter used for calculating the onset probability using, for example, a logistic model, and stores the parameter in the model storage unit 14.

  In addition, the risk evaluation unit 15 uses the model in the model storage unit 14 and the medical examination information in the medical examination information storage unit 11 after the predetermined period described above (for example, for the latest year). The probability of onset is calculated. The calculated onset probability is stored in the result storage unit 16 in association with the identification information of the subject.

  In the present embodiment, the model generation unit 13 generates a model for calculating an onset probability using a medical examination result at a specific time point within a predetermined period and a medical cost spent within the predetermined period. That is, for example, it is possible to learn the characteristics of the medical examination information and the number of visits at the time when the medical cost is more than a predetermined amount. If the onset probability is calculated using such a model, the risk assessment unit 15 can evaluate the current risk level of the target person from the viewpoint of the overall medical cost reduction effect, The target person can be extracted based on the size.

<Device configuration>
FIG. 2 is an apparatus configuration diagram illustrating an example of a computer. The risk analysis apparatus 1 is a computer as shown in FIG. 2 includes a CPU (Central Processing Unit) 1001, a main storage device 1002, an auxiliary storage device 1003, a communication IF (Interface) 1004, an input / output IF (Interface) 1005, a drive device 1006, and a communication bus 1007. ing. The CPU 1001 performs processing according to the present embodiment by executing a program (also referred to as “software” or “application”). The main storage device 1002 caches programs and data read by the CPU 1001 and develops a work area of the CPU. Specifically, the main storage device is a RAM (Random Access Memory), a ROM (Read Only Memory), or the like. The auxiliary storage device 1003 stores programs executed by the CPU 1001, setting information used in the present embodiment, and the like. Specifically, the auxiliary storage device 1003 is an HDD (Hard-disk Drive), an SSD (Solid State Drive), a flash memory, or the like. The main storage device 1002 and the auxiliary storage device 1003 function as the medical examination information storage unit 11, the reception storage unit 12, the model storage unit 14, and the result storage unit 16. For convenience of explanation, FIG. 1 shows a plurality of storage units (the medical examination information storage unit 11, the receipt storage unit 12, the model storage unit 14, and the result storage unit 16). A plurality of storage devices may be used. The communication IF 1004 transmits / receives data to / from other computers. The risk analysis apparatus 1 may receive medical examination information and receipt information from a computer (not shown) connected via the communication IF 1004. The communication IF 1004 is specifically a wired or wireless network card or the like. The input / output IF 1005 is connected to the input / output device and accepts input from the user or outputs information to the user. Specifically, the input / output device is a keyboard, a mouse, a display, a touch panel, or the like. The drive device 1006 reads data recorded on a storage medium such as a magnetic disk, a magneto-optical disk, and an optical disk, and writes data to the storage medium. The above components are connected by a communication bus 1007. A plurality of these components may be provided, or some of the components (for example, the drive device 1006) may not be provided. Further, the input / output device may be integrated with the computer. In addition, the program executed in this embodiment is provided via a portable storage medium readable by the drive device 1006, a portable auxiliary storage device 1003 such as a flash memory, a communication IF 1004, and the like. It may be. When the CPU 1001 executes the program, the computer as described above is operated as the risk analysis apparatus 1 shown in FIG.

  Note that at least a part of the functional configuration illustrated in FIG. 1 may exist on the network. For example, a so-called cloud service by one or more computers having at least a part of the configuration of FIG. 1 may be provided.

<Risk analysis processing>
FIG. 3 is a process flow diagram illustrating an example of the risk analysis process. It is assumed that the medical examination information storage unit 11 and the receipt storage unit 12 hold, for example, the results of a medical examination of a subject belonging to a predetermined health insurance association and detailed information on medical expenses, respectively. In the risk analysis process, first, the model generation unit 13 of the risk analysis device 1 generates a model represented as a function for calculating a predetermined onset probability (FIG. 3: S1). For convenience, although referred to as “onset probability”, in this step, a model is generated for calculating the probability that the medical cost will be greater than or equal to a predetermined amount for a predetermined injury and disease name in a predetermined period.

  FIG. 4 is a process flow diagram illustrating an example of the model generation process. The model generation unit 13 reads a record of a medical checkup that the subject has received during the predetermined period from the medical checkup information storage unit 11 (FIG. 4: S11). In this step, the model generation unit 13 acquires a medical examination history including measurement values of predetermined examination items from the medical examination information. For example, when paying attention to a disease such as a predetermined brain, heart, or circulatory organ as a disease that affects medical expenses, the model generation unit 13 determines blood pressure (systole and diastole), lipid ( HDL, LDL, TG), blood glucose level, and health checkup history of HbA1c are acquired. In this step, for example, medical examination information for a period in which the amount of medical expenses for one year, which will be described later, is fixed, such as up to the previous year at the time of processing, is acquired.

  Moreover, the model production | generation part 13 reads the receipt information which generate | occur | produced in the subject in the predetermined period from the receipt memory | storage part 12, and totals it (S12). In this step, the name of the main injury and the medical cost are acquired as the receipt information in association with the identification information of the subject. It is assumed that the identification information of the target person is the same as or known to correspond to the identification information of the target person in the above-described medical examination information. Note that the order of the process of S11 and the process of S12 may be interchanged or may be executed in parallel. In addition, the medical examination information and the receipt information are used for a predetermined number of years over the past one year.

As main injury and illness names used in S12, for example, hypertension, hyperlipidemia, diabetes and the like can be used. The main wound name is the number of visits (medical visits) related to hypertension and medical expenses, the main wound name is the number of visits and medical expenses related to hyperlipidemia, the main wound name is the number of visits and medical expenses related to diabetes, and the main wound name is The number of consultations related to a disease such as a predetermined brain, heart or circulatory organs shall be counted. Specifically, medical examinations related to hypertension are the names of diseases and diseases having the same meaning as those classified by the International Classification of Disease (hereinafter referred to as ICD), etc. It is also possible to refer to a receipt that includes these names. Medical examinations related to hyperlipidemia may include the names “high fat” or “lipid” or “cholesterol” in the disease name, or synonymous injury and disease names classified by ICD, etc. Refers to a receipt that includes a name. In addition, a diagnosis regarding diabetes may be a name containing the word “sugar” in the disease name, or a synonymous wound name classified by ICD or the like, and refers to a receipt including these wound names. In addition, the diagnosis related to a disease such as a predetermined brain, heart or circulatory organ is defined as “stroke”, “subarachnoid hemorrhage”, “intracerebral hemorrhage”, “cerebral arteriosclerosis”, “cerebral embolism”, “cerebral vascular disorder” ”,“ Cerebral hemorrhage ”,“ intracerebral hemorrhage ”,“ cerebral thrombus ”,“ cerebrovascular disorder ”,“ brain stem hemorrhage ”,“ intraventricular hemorrhage ”,“ subcortical hemorrhage ”,“ cerebral artery circulation ”,“ cerebral circulation failure ” ”,“ Subdural hematoma ”,“ subcortical hemorrhage ”,“ intracranial hemorrhage ”,“ cardiomyopathy ”,“ angina ”,“ aneurysm ”,“ arterial dissection ”,“ arterial stenosis ”,“ arterial occlusion ” ”,“ Arterial closure ”,“ arterial thrombus ”,“ arteriovenous fistula ”,“ venous embolism ”,“ varicose vein ”,“ venous fistula ”,“ ischemia ”,“ infarction ”(excluding traumatic cases) ), Synonymous wound names classified by ICD, etc., other renal diseases and liver diseases, including these wound names It is assumed that refers to the concept. For example, medical expenses paid by the name of the main injury or illness for a disease such as a predetermined brain, heart, or circulatory organ to be described later are totaled.

  Next, the model generation unit 13 converts the acquired measurement value and the collected receipt information into a predetermined score (S13). In this step, based on a predetermined condition, the predetermined measurement value and the number of visits (visiting hospitals) with a predetermined injury and illness are converted into scores used as explanatory variables in the statistical analysis according to the present embodiment.

FIG. 5 is a diagram illustrating an example of conditions used for conversion to a score. When paying attention to a disease such as a predetermined brain, heart or circulatory organ, for example, such a condition can be adopted. X ₁
Is a score (0 to 2 points) related to high blood pressure, and is determined based on a measured value of blood pressure (at the time of contraction and expansion) and a medical examination history with the name of the disease. In S13, when the condition registered in the field of the “condition” column is satisfied, the model generation unit 13 applies the score associated with the condition. When a plurality of conditions are satisfied, a score having a large value is applied.

Specifically, if the record of the medical examination regarding hypertension described above is not included in the receipt information, and the condition of blood pressure (expansion) <100 and blood pressure (contraction) <160 of the medical examination information is satisfied, the score of X ₁ is 0. . Further, if the number of consultations related to hypertension is 1 to 4 or if the condition 100 ≦ blood pressure (dilation) <110 or 160 ≦ blood pressure (contraction) <180 is satisfied, the score of X ₁ is 1. If the number of visits related to high blood pressure is 5 or more, or the condition 110 ≦ blood pressure (dilation) <120 or 180 ≦ blood pressure (contraction) <200 is satisfied, the score of X ₁ is 2
And When a plurality of conditions are met, the higher score is preferentially applied.

X ₂ is the score (0 to 2 points) for hyperlipidemia and lipid (HDL, LDL, TG)
It is determined based on the measured value and the history of visits with the name of the disease. Specifically, if there is no record of medical examination regarding hyperlipidemia in the receipt information and 40 ≦ HDL and LDL ≦ 140 in the medical examination information, the score of X ₂ is 0. Further, when the number of consultations related to hyperlipidemia is 1 to 4 in the reception, or in the medical examination information, HDL <40 or 140 ≦ LDL <180, the score of X ₂ is 1. Also, if the number of consultations related to hyperlipidemia is 5 or more at the reception, or 300 ≦ TG or 180 ≦ LDL in the medical examination information, the score of X ₂ is 2. When a plurality of conditions are met, the higher score is preferentially applied.

X ₃ is a score (0 to 2 points) related to diabetes, and is determined based on the blood glucose level, the value of HbA1c, and the visit history with the name of the disease. Specifically, if there is no record of medical examination regarding blood glucose level in the receipt information and the blood glucose level <130 and HbA1c <6.5 in the medical examination information, the score of X ₃ is set to 0. Further, if the number of consultations related to diabetes is 1 to 4 in the reception, or 130 ≦ blood glucose level <160 or 6.5 ≦ HbA1c <8.0 in the medical examination information, the score of X ₃ is 1. Further, if the number of consultations related to diabetes is 5 or more at the reception, or 160 ≦ blood glucose level or 8.0 ≦ HbA1c in the medical examination information, the score of X ₃ is 2. When a plurality of conditions are met, the higher score is preferentially applied.

X ₄ is a score (0 to 2 points) related to a disease such as a predetermined brain, heart or circulatory organ,
It is determined on the basis of the number of times of consultation with the name of a disease related to a disease such as a predetermined brain, heart or circulatory organ. Specifically, given the brain, a record of visits related disorders such as heart or circulatory not in the receipt information, the score of X ₄ is 0. In addition, when the number of consultations regarding a disease such as a predetermined brain, heart, or circulatory organ is 1 to _{4 in} the reception, the score of X ₄ is 1.
In addition, if the number of medical examinations related to a disease such as a predetermined brain, heart, or circulatory organ is 5 or more at the reception, the score of X ₄ is 2.

As described above, in this embodiment, instead of directly using the measurement value or the number of medical examinations, it can be converted into a score at a predetermined stage (that is, an ordinal scale) to be an explanatory variable. That is, the scores of X _{1 to} X ₄ can be said to be evaluation values related to a predetermined injury and illness based on the result of the health check included in the medical checkup information and the status of the hospital visit included in the receipt information. The inventors have found that these conditions used for score determination are suitable for explanatory variables for calculating the onset probability described later. A part of these scores may be used, or a score not illustrated may be further used. The selection of such medical examination items and the conditions for conversion to scores can be automatically determined by the computer. The score of S13 is obtained for each of a plurality of subjects. These evaluation values can be used by appropriately setting variables after X ₅ as necessary.

And the model production | generation part 13 produces | generates the model for calculating a predetermined onset probability based on the medical examination information and the receipt information of a plurality of subjects (S14). Any linear function or non-linear function can be used as long as the object of the present invention is a method that can achieve an appropriate evaluation of the current risk level of the subject. In this case, the model may be a logistic function for obtaining a probability that the medical cost is equal to or higher than a predetermined threshold value with a predetermined injury / illness name. In this step, the onset probability P is expressed by the following logistic function (formula (1)).
Logit (logarithmic odds) Z is expressed by the following logit model (formula (2)) using the above-mentioned score.
Z = β ₀ + β ₁ · X ₁ + β ₂ · X ₂ + β ₃ · X ₃ + β ₄ · X ₄ (2)

In this step, five parameters β _{0 to} β ₄ are obtained by the likelihood method. That is, assuming that the above-mentioned onset probability of the j-th subject is P _j , a disease such as a predetermined brain, heart or circulatory organ is the name of the main injury and a predetermined period of medical expenses (for example, 1 million yen per year) If this is the case,
L _j = P _j
Otherwise,
L _j = 1−P _j
far. Then, parameters β _{0 to} β ₄ that maximize the following likelihood function (equation (3)) are obtained.
like (β) = ln (L ₁ × L ₂ × ... L _j × ...) = Σln (L _j ) (3)
Note that the parameters β _{0 to} β ₄ with the maximum likelihood function can be obtained by using an existing technique such as a hill-climbing method.

The parameters β ₀ to β ₄ are determined as described above, and the medical cost is more than a predetermined amount based on a predetermined score related to hypertension, hyperlipidemia, diabetes and a disease such as a predetermined brain, heart, or circulatory system. A function for obtaining the onset probability P of a disease such as a predetermined brain, heart, or circulatory organ can be generated.

  Thereafter, returning to the process of FIG. 3, the risk evaluation unit 15 performs a risk evaluation process (FIG. 3: S2).

  FIG. 6 is a process flowchart showing an example of the risk evaluation process. First, the risk evaluation unit 15 acquires the latest medical examination information of the subject and the receipt information for the most recent predetermined period (FIG. 6: S21). In this step, the risk evaluation unit 15 includes, for example, predetermined test values and latest data related to the next year of the medical examination information acquired in S11 among the medical examination information of the subject stored in the medical examination information storage unit 11. Extract health checkup information for one year. Specifically, the risk evaluation unit 15 acquires blood pressure (systolic and diastolic), lipid (HDL, LDL, TG), blood glucose level, and measurement values of HbA1c health checkup. In addition, the latest health risk of the target person at the time of processing can be evaluated by using the latest medical examination information, but the health risk assumed at that time is evaluated using the medical examination information at any time You can also

Next, the risk evaluation unit 15 calculates the onset probability of the subject using the function of the onset probability P generated in S14 (S22). In this step, the risk evaluation unit 15 first converts the health check measurement value and the visit history acquired in S21 into a predetermined score based on the same conditions as in S12. Then, the onset probability P shown in Expression 1 is calculated using the parameters β _{0 to} β ₄ generated in S14.

And the risk evaluation part 15 extracts the object person of health guidance (S23). For example, the risk evaluation unit 15 extracts a predetermined upper number of subjects based on the magnitude of the onset probability P, that is, as subjects with a high risk. The parameters β _{0 to} β ₄ are parameters obtained on the basis of the measured value of the health check and the medical expenses for the year of receiving the health check, and the onset probability P calculated in S22 is exactly the time at the time of the processing. It can be said that it represents the risk of the target person. Therefore, in S23, it is possible to extract a high-priority target person who performs health guidance, which is efficient. In addition, when health guidance is given based on the amount of medical expenses in the current year, if the most recent period is specified as the specified period and the data is used, the target person receiving the health guidance will be more urgent rather than several years later. In addition, since it is possible to recognize that this is a highly likely risk, it is easy to promote health guidance such as improvement of lifestyle habits, which is preferable.

In addition, the person who gives the health guidance and the target person to be instructed can easily explain and understand the basis of the risk if the risk can be calculated by a simple method in the place of health guidance, for example. For example, the logit of this embodiment includes a measurement value and a visit history related to hypertension, a measurement value and a visit history related to hyperlipidemia, a measurement value and a visit history related to diabetes, and a predetermined brain, heart, or circulatory organ. It is a linear expression using the visit history regarding diseases such as, and the simplicity is also an advantage when used for health guidance. However, the present invention can also use a non-linear function to represent the onset probability.

In addition, when the target person is a specific group such as a person belonging to a predetermined health insurance association, the target person may have a biased characteristic compared to a general standard. That is, there is a case where a certain population of subjects has a lower probability of occurrence of a predetermined injury or illness than a general standard. In such a case, when there is a limit to the human resources for providing health guidance to the target group, it is possible to extract high-risk target persons with high accuracy. It is efficient to provide health guidance to high-risk subjects. Therefore, for example, based on medical examination information and receipt information of subjects belonging to a predetermined group such as subjects belonging to a predetermined health insurance association, subjects working for a predetermined company, subjects of a predetermined industry or occupation, etc. It can be said that the accuracy of risk evaluation is improved by using the generated onset probability model when determining the risk level of subjects belonging to the same group. Thus, by using the present invention, it is possible to appropriately grasp the current state of health of the subject and the magnitude of the risk. Therefore, it becomes easier to decide the priority of the response such as health guidance, and it can be implemented so that the effect can be maximized within the limited resources. As a result, it will contribute to the reduction of medical expenses. preferable.
Moreover, since it becomes possible for the subject himself / herself to grasp the health state before getting sick, this leads to satisfying the desire to extend a healthy life span and live a full and fulfilling life for a long time.
In addition, based on the evaluation made according to the present invention, it can be sufficiently effective in the development of consulting corresponding to each target person such as appropriate nutritional guidance, exercise guidance, and selection of functional foods. It becomes possible to do so.

<Update probability of onset>
Later, when the medical expenses of the subject in the year in which the above-described processing is performed are confirmed, the onset probability P can be updated using a combination of the medical examination information and the receipt information of the latest year. That is, the above-described model creation process is executed by further using the medical examination information and the receipt information for the new year. In this way, the risk at that time can be more accurately evaluated based on the onset probability model reflecting the latest tendency of the subjects belonging to the group. The receipt information may be updated every time the patient visits the hospital, and the onset probability P may be obtained using the medical examination information and the receipt information for the most recent predetermined year.

<Example 1>
The upper part of FIG. 7 calculates the onset probability P based on the medical examination information and receipt data for 31367 people in a certain year (first year) and the following year (second year), and the order of increasing probability P It is the graph which arranged the subject's onset probability from right to left. The area painted in black indicates the probability of onset. Specifically, the vertical length represents the magnitude of the individual onset probability, and the person with the highest onset probability is 27%. The lower part of FIG. 7 is a graph obtained by enlarging the top 2000 people in the upper graph. When generating the onset probability P, which is the model shown in the formula (1), a score related to hypertension, a score related to hyperlipidemia, a score related to diabetes shown in FIG. 5, a predetermined brain, heart, circulatory organ, etc. Using the disease score, it was defined that the disease occurred when the annual medical cost exceeded 500,000 yen due to a disease such as a predetermined brain, heart, or circulatory organ. For reference, subjects whose annual medical expenses exceeded 500,000 yen are indicated by broken lines on the drawing. The short dashed line (127 people) indicates subjects whose medical expenses exceeded 500,000 yen in the first and second years, and the long broken line (96 persons) shows subjects whose medical expenses exceeded 500,000 yen in the third year Shows the person. In addition, subjects whose medical expenses exceeded 500,000 yen for both periods are also included.

That is, FIG. 7 is a graph in which the subject is prioritized by the onset probability calculated based on the medical examination information and the receipt information for two years. Referring to the lower graph, among the top 2000 people, at the time of calculating the probability of onset (at the end of the second year), they did not pay more than 500,000 yen, but in the third year, 500,000 yen The target person who paid more medical expenses
Shown with long dashed lines. In the first to third years, subjects who have not paid medical expenses exceeding 500,000 yen but are judged to be at high risk are shown without a broken line in the lower part of FIG. At the end of the second year, providing health guidance to subjects shown with long dashed lines or subjects without dashed lines efficiently allocates human resources to health guidance for high-risk subjects be able to. In addition, according to such ranking, for example, unlike a case where a subject is high in blood sugar level, and a subject is high in cholesterol value, the individual ranking is focused on only specific test results. It is possible to save the labor of the instructor in judging the results in a complex manner. If the data from the third year onward are further used, it is expected that the accuracy of the onset probability is further improved.

<Example 2>
The upper part of FIG. 8 is a diagram in which onset probabilities are calculated based on the medical examination data and receipt data for 31 years of 31367 subjects and are arranged from right to left in descending order of onset probability P. The lower part of FIG. 8 is an enlarged display of about 2500 people with the highest probability of onset in the upper part of FIG. Moreover, the subject whose medical expenses exceeded 500,000 yen in any one of three years was shown with the broken line in FIG. When generating the onset probability P, which is the model shown in the formula (1), a score related to hypertension, a score related to hyperlipidemia, a score related to diabetes shown in FIG. 5, a predetermined brain, heart, circulatory organ, etc. Using the disease score, it was defined that the disease occurred when the annual medical cost exceeded 500,000 yen due to a disease such as a predetermined brain, heart, or circulatory organ.

When the parameters β _{0 to} β ₄ in Equation (2) are obtained using the scores shown in FIG. 5, the subject who has paid a relatively high medical cost as shown in FIG. 8 (exists at the position indicated by the broken line) Can be extracted with high precision (upper right). In addition, in a position where a broken line is not attached, a medical person with a high risk is extracted at the time of the ranking generation process, although a medical fee of a predetermined amount or more has not yet occurred. In particular, it is possible to cope with such a subject by providing health guidance and the like, and human resources such as health guidance can be efficiently allocated to a high-risk subject.

<Comparative example>
FIG. 9 is a diagram arranged from the right to the left in descending order of the total value based on the value of HbA1c for 31367 subjects for three years. In addition, the subject whose medical expenses exceeded 500,000 yen in any one of three years was shown with the broken line in FIG. FIG. 10 is a diagram in which points are assigned on the basis of blood glucose, blood pressure, and lipid values for 31367 subjects for three years and arranged from right to left in descending order of the total score. In addition, the blood glucose was set to 1 point when HbA1c ≧ 6.5 or fasting blood glucose level ≧ 126. The blood pressure was 1 point when systolic blood pressure ≧ 160 or diastolic blood pressure ≧ 100. Lipids were scored at 1 point for LDL ≧ 180 or TG ≧ 1000. In addition, a subject whose medical expenses exceeded 500,000 yen in any one of the three years is also indicated by a broken line in FIG.

  Compared with the results shown in FIG. 8, when the medical items are sorted according to the single item values shown in FIG. 9 or sorted according to the scores corresponding to the plurality of items shown in FIG. Can't get a higher rank. Temporarily, there are 1500 target persons who can give health guidance in terms of resources, and the top 1500 persons who have evaluated the risk by each method are targeted for health guidance. In this case, subtracting the medical cost of the top 1500 people in FIG. 9 from the medical cost of the top 1500 people in FIG. 8 will be about 400 million yen / year. Also, if the medical expenses of the top 1500 people in FIG. 10 are subtracted from the medical expenses of the top 1500 people in FIG. 9, it will be about 500 million yen / year. In other words, if it can be prevented even partly by health guidance, it can be said that it will lead to a reduction in medical expenses of 400 to 500 million yen per year.

<Modification of Embodiment>
It is known that clinical laboratory values may have a certain degree of variation depending on the laboratory or reagent used. In order to calculate the risk evaluation value without reflecting such an error between the measurement values, a correction coefficient representing a deviation from the measurement result of the standard substance can be used. That is, the influence of the error can be reduced by multiplying the measured value by the correction coefficient. Such correction is preferable because risk can be accurately evaluated.

In addition, the model generation unit 13 may use the acquired measurement values and the collected receipt information as they are as explanatory variables without converting them into a predetermined score. That is, without executing S13 in FIG. 4, for example, substituting the measured values and visits count to X ₁ to X ₄ of the above-mentioned equation (2). For example, in S14, blood pressure, lipid, blood glucose level, HbA1c, and other medical examination counts, which are measurement values used for the above-described logistic function, are measured. In addition, you may use the frequency | count of consultation itself regarding the four types of diseases mentioned above as an explanatory variable. At this time, it is particularly preferable to use the above-described correction coefficient because the influence of the error of the explanatory variable on the risk evaluation can be suppressed.

  When using the acquired measured value or the aggregated value of the receipt information as an explanatory variable, it is not necessary to set the standard compared to the case where the measured value in each test is scored under a certain standard. It can be carried out more simply. Further, since the value is used as it is without being converted, the calculation result is in line with the actual situation, which is preferable. For example, it is possible to legitimately assess the risk of a patient who is within the criteria but is transitioning in the border region. In addition, when scoring in several stages, the number of subjects who have the same onset probability is likely to increase. On the other hand, by using it as an explanatory variable without scoring, it becomes easy to give a difference in rank to the target person. That is, for example, when a predetermined number of high-order subjects are extracted, there is a high possibility of being divided at a position close to a desired number of people.

<Example 3>
As the explanatory variable described above, the examination item value and the number of consultations may be used as they are. Eleven variables of systolic blood pressure, diastolic blood pressure, neutral fat, HDL, LDL, blood glucose level, HbA1c test value, diabetes, hyperlipidemia, hypertension, and number of visits for cerebrocardiovascular disease are used as explanatory variables as they are. When the onset probability of cerebrocardiovascular disease was evaluated by regression, the ROC (Receiver Operating Characteristic) curve of FIG. 11 was obtained. The ROC curve is a graph in which the vertical axis represents the true positive ratio, the horizontal axis represents the false positive ratio, and the coordinates corresponding to the test item values are plotted. The larger the area under the curve (AUC), the higher the discrimination performance. For comparison, when the onset probability is evaluated using only the LDL cholesterol value considered to be related to arteriosclerosis as an explanatory variable and an ROC curve is drawn, FIG. 12 is obtained. In this way, high discrimination performance is obtained in the present invention by combining the medical examination data and the number of medical examinations.

<Others>
As mentioned above, the various constituent elements and evaluation conditions of the present invention have been illustrated, but the scope of the present invention is not limited thereto, and those skilled in the art can appropriately select them with reference to the above conditions. is there. In particular, the types of explanatory variables used to determine the onset probability are not limited to those described in the embodiment, Examples 1 to 3, and modifications, and may be appropriately selected according to, for example, the injured disease. Can do.

  The present invention includes a computer program that executes the above-described processing and a computer-readable recording medium that records the program. The recording medium on which the program is recorded can perform the above-described processing by causing the computer to execute the program.

Here, the computer-readable recording medium refers to a recording medium in which information such as data and programs is accumulated by electrical, magnetic, optical, mechanical, or chemical action and can be read from a computer. Examples of such a recording medium that can be removed from the computer include a flexible disk, a magneto-optical disk, an optical disk, a magnetic tape, and a memory card. In addition, examples of the recording medium fixed to the computer include an HDD, an SSD (Solid State Drive), and a ROM.

DESCRIPTION OF SYMBOLS 1 Risk analyzer 11 Medical examination information storage part 12 Receptacle storage part 13 Model generation part 14 Model storage part 15 Risk evaluation part 16 Result storage part

Claims (9)

Obtaining from the storage unit medical checkup information including measurement values of medical examinations in a predetermined period of a plurality of subjects, and receipt information including medical expenses and names of sicknesses in the predetermined period of the plurality of subjects;
Generating a model representing a probability that a medical cost is equal to or higher than a predetermined threshold in the reception information with a value based on the medical examination information and the reception information as an explanatory variable;
Reading the medical examination information of the subject after the predetermined period from the storage unit, calculating a risk evaluation value using a value based on the medical examination information and the model;
A risk assessment method that the computer performs. The value based on the medical examination information and the receipt information is a value obtained by converting one or more measured values in the medical examination and the number of medical examinations with a predetermined injury / illness name in the reception information into a score represented by a predetermined order scale. The risk evaluation method according to claim 1. The risk evaluation method according to claim 1, wherein the value based on the medical examination information and the receipt information is a measurement value of a predetermined medical examination item in the medical examination information and the number of medical examinations with a predetermined injury / illness name in the reception information. 4. The value based on the medical examination information includes at least one of a measurement value related to hypertension, a measurement value related to hyperlipidemia, or a measurement value related to diabetes. 5. Risk assessment method described in 1. The risk evaluation method according to any one of claims 1 to 4, wherein the model is represented by a logistic function for obtaining the probability that the medical cost is equal to or greater than a predetermined threshold value with a predetermined injury / illness name. The risk evaluation value is calculated for a plurality of subjects, and the computer further executes a step of extracting a predetermined upper number of subjects based on the magnitude of the calculated risk assessment value. The risk evaluation method according to claim 1. When the medical examination information and the receipt information in a period after the predetermined period are registered in the storage unit, the computer further executes the step of generating the model by further using the medical examination information and the receipt information in the period The risk evaluation method according to any one of claims 1 to 5. A storage unit including medical examination information including measurement values of medical examinations in a predetermined period of a plurality of subjects, and receipt information including medical expenses and names of sicknesses in the predetermined period of the plurality of subjects,
The medical examination information and the receipt information are acquired from the storage unit, and a value based on the medical examination information and the receipt information is used as an explanatory variable. A model generation unit for generating a model representing the probability of
The risk assessment unit that reads the medical examination information of the subject after the predetermined period from the storage unit, and calculates a risk evaluation value using the value based on the medical examination information and the model,
A risk evaluation apparatus comprising: Obtaining from the storage unit medical checkup information including measurement values of medical examinations in a predetermined period of a plurality of subjects, and receipt information including medical expenses and names of sicknesses in the predetermined period of the plurality of subjects;
Generating a model representing a probability that a medical cost is equal to or higher than a predetermined threshold in the reception information with a value based on the medical examination information and the reception information as an explanatory variable;
Reading the medical examination information of the subject after the predetermined period from the storage unit, calculating a risk evaluation value using a value based on the medical examination information and the model;
Risk assessment program that runs on a computer.