JP4662509B1 - Cognitive function prediction system - Google Patents

Abstract

A system for predicting future cognitive functions of individuals with onset of dementia is provided.
A cognitive function prediction system 1 of the present invention includes an inspection server 10 and an inspection terminal 30 connected via a communication line 50, and the inspection server 10 includes past cognitive functions for a large number of subjects. By the non-linear regression analysis using the test score in the test and the test time from the onset to the test date, a prediction calculation formula for predicting the future test score of each target subject is estimated, and the test terminal 30 The prediction calculation formula is executed by substituting the past test result and the current test result of the subject subject into the obtained prediction formula, and the future test score of the target subject is generated.
[Selection] Figure 2

Description

  The present invention relates to a cognitive function prediction system that predicts a future cognitive function of an individual with dementia.

  As life expectancy increases and the number of the elderly increases, care for the elderly, especially dementia such as Alzheimer-type dementia, vascular dementia, and Lewy body dementia, which is now said to exceed 2 million. The problem of nursing care for elderly people who have suffered from this has become socially important.

  Although dementia is considered to be basically uncurable, appropriate use of acetylcholinesterase inhibitors such as donepezil hydrochloride (trade name Aricept (registered trademark), manufactured by Eisai Co., Ltd.) and other pharmaceuticals Or by appropriate rehabilitation and training, it is possible to delay the progress of dementia and prolong the time that the person with dementia develops and the time that the person with dementia presents with the family. is there. Therefore, when dementia is suspected, it is important to receive an appropriate cognitive function evaluation as early as possible.

By the way, in order to obtain knowledge such as whether dementia has developed, what level of dementia has developed, or how much dementia has progressed, the following Table 1 Various types of cognitive function tests have been proposed, and have been performed by medical personnel such as doctors, clinical psychologists, and speech auditors.

  For example, the Japanese version of Alzheimer's Disease Assessment Scale shown in Non-Patent Document 1 (hereinafter referred to as “ADAS-Jcog”) is intended to monitor the cognitive function of persons with Alzheimer-type dementia. It is a cognitive function test.

  There has also been proposed a cognitive function test system that performs a cognitive function test on a person with dementia or who is suspected of developing dementia using a computer.

  For example, Patent Literature 1 discloses a test item designating unit that designates one of a plurality of intelligent test items according to an input operation, an intelligent test problem corresponding to the intelligent test item designated by the test item designating unit, and its answer. And a score evaluation means for scoring with reference to the answer according to the answer operation for the intelligence test problem and generating an evaluation result by scoring the score. As the applicable intelligence tests, there are dementia scales for the elderly such as “Hasegawa-style simplified intelligence assessment scale” and “National Seken-style dementia screening test”, questionnaire-style behavior assessment scales and psychiatric symptom assessment scales. Illustrated. With this test apparatus, the dementia tendency can be easily determined. This document further describes that a score storage means for storing the total score in time series is provided, and the past score read from the score storage means is compared with the current score for relative evaluation.

  Patent Document 2 also includes an answer acquisition unit that obtains a subject's answer to a dementia degree test chart that simultaneously obtains a plurality of types of judgments and obtains answers in a format in which judgment correctness is objectively determined. There is disclosed a dementia test apparatus including a dementia degree test unit that tests a dementia degree of a subject based on an answer obtained by an answer acquisition unit. The dementia degree test chart is used to determine a decrease in frontal lobe function that causes senile dementia. The test results of a plurality of subjects are transmitted to a data storage server via a communication line, and the transmitted test results are monitored by the subjects and statistical processing on a large number of subjects (implemented) In the form, it is used for processing for obtaining an average value and variance for each age).

  Patent Document 2 further provides a dementia degree test chart for obtaining answers in a format in which multiple types of judgments are simultaneously requested and whether the judgment is objectively judged, and a plurality of questions relating to sensibility and each question are prepared. An answer acquisition unit for acquiring a subject's response to both the dementia factor degree test chart composed of a combination of a plurality of alternatively selected responses, and a response based on the answer obtained by the response acquisition unit. There is described a dementia test apparatus including a dementia degree test unit that performs a test of a dementia degree indicating the current degree of dementia of the examiner and a transition prediction of the future dementia degree of the subject. The dementia factor degree test chart focuses on the fact that the activity of the right brain becomes dominant as it ages and tests how the right brain that controls sensibility is used on a daily basis (not used) For each group classified as “married man”, “full-time housewife”, etc., it is evaluated how much the brain is activated in “daily life and workplace”, “daily life and housework”, and the like. In the embodiment, the horizontal axis is the score of the frontal lobe function test using the dementia degree test chart, the vertical axis is the score of the right brain activity test using the dementia factor degree test chart, and each age (20s, 30s,... , 100s) In a graph showing the least square approximation line of the score distribution of a large number of subjects in the past, the test result of a subject is shown on the least square approximation line in the age to which this subject belongs As a result, the current degree of dementia of the subject is tested, and it is further assumed that the score of the right brain activity test does not change after 10 or 20 years. By reading the score of the frontal lobe function test corresponding to the score of the current right brain activity test on the least square approximation line, the transition prediction of the future dementia degree of the subject is performed.

JP-A-9-16072 JP 2002-143096 A

Geriatric Psychiatric Journal, 3 (6), 647-655, (1992)

  Symptoms from the comparison of scores to the present when using a device that can compare past scores and current scores in a subject's cognitive function test as shown in Patent Literature 1 and Patent Literature 2 It is possible to determine the degree of decrease or recovery of the drug, and for example, it can be used to determine the effectiveness of the medicine taken up to now. However, even if this subject's past score is compared with the current score, the future cognitive function cannot be predicted. On the other hand, if the future cognitive function of the subject can be predicted, effective responses in various fields such as medical care, care, and life planning according to the future state change of the subject can be examined. In addition, because it can encourage the preparation of the subject and the caregiver's family, the time that the person with dementia develops his / her personality, the time that the person with dementia is with the family It can be enriched. Furthermore, if the difference in the future cognitive function depending on the presence or absence of medical treatment such as medication can be predicted, it is possible to know the predicted effectiveness of the medical treatment, which can lead to more effective medical treatment.

  Although Patent Document 2 shows a device that predicts the transition of the future dementia level of a subject, the prediction of the future cognitive function is based on the assumption that the score of the right brain activity test does not change in the future, Changes in cognitive function over time from the onset of dementia in persons with dementia are not considered. In addition, the validity of the assumption that the score of the right brain activity test will not change in the future is suspected. Furthermore, with this device, it is difficult to make future predictions using information specific to the subject, such as the decline in cognitive function of the subject to date or the elapsed time from the onset of dementia to treatment. .

  Therefore, an object of the present invention is to predict future cognitive function more accurately for each person with dementia and to predict future cognitive function using information unique to the subject. It is to provide a cognitive function prediction system that can also.

  The cognitive function prediction system of the present invention that achieves the above object includes a test server and a test terminal connected via a communication line, and the test terminal is a dementia for a subject who has developed dementia. A test result acquisition unit that acquires an onset time and a test date and a test score in a cognitive function test, a test subject's dementia onset time acquired in the test result acquisition unit, a test date and a test score in a cognitive function test, A test result transmission unit that transmits the test result to the test server, and the test server receives the dementia onset time of the subject and the test date and test score in the cognitive function test transmitted from the test terminal. A test result receiving unit, and a test result storage unit that stores for each subject the onset date of dementia and the test date and test score in the cognitive function test for a large number of subjects. Inspection end Is a cognitive function prediction system for predicting the future test score of the subject subject whose current test score has been acquired by the test result acquisition unit, wherein the test server is stored in the test result storage unit. For predicting future test scores of the above-mentioned subjects by non-linear regression analysis using test scores in the past cognitive function tests and the elapsed time from the onset to the test date A prediction calculation formula estimating unit that estimates a prediction calculation formula, a prediction calculation formula storage unit that stores a prediction calculation formula estimated by the prediction calculation formula estimation unit, and a prediction calculation formula stored in the prediction calculation formula storage unit A prediction data transmission unit that transmits to the inspection terminal, and the inspection terminal receives the prediction calculation expression transmitted from the inspection server, and the prediction data reception unit receives the prediction calculation expression. Forecast A prediction result generation unit that generates the future test score of the target subject by applying the current test score of the target subject to the arithmetic expression, and the target subject generated by the prediction result generation unit And a prediction result display unit for displaying a future examination score.

  “Target subject” means an individual subject to be predicted for future cognitive function, and in the cognitive function prediction system of the present invention, the current cognitive function test for the target subject is completed. Immediately after, the future cognitive function of the subject subject is predicted. The future cognitive function of the target subject is predicted from the test score of the cognitive function test of many subjects and the elapsed time from the onset to the test date. The number of subjects providing data for nonlinear regression analysis, that is, the meaning of `` many '' should be sufficient for nonlinear regression analysis. This is preferable because a good prediction formula is estimated. Since the system of the present invention is a so-called server-client system, it is possible to quickly collect data of a large number of subjects by connecting an inspection server and a large number of inspection terminals via a communication line. Therefore, the accuracy of prediction of the future cognitive function of the subject subject can be quickly improved. And in the system of this invention, since the future cognitive function is predicted with reference to the transition of the cognitive function in persons with dementia, compared with the device of Patent Document 2, the prediction of the future cognitive function is more accurate. It becomes possible.

  The non-linear regression analysis for estimating the prediction formula may be performed on the data of all the subjects stored in the test result storage unit of the test server. It can also be done as a target. In particular, as a result of analyzing the transition of cognitive function for a large number of subjects, for example, a large difference was observed between subjects with a history of cerebrovascular disease and subjects without a history of cerebrovascular disorders. In some cases, if nonlinear regression analysis is performed on data of a subject who has the same history of cerebrovascular disorders as the subject, the future cognitive function of the subject can be predicted more accurately. it can.

  Therefore, in a preferred form of the cognitive function prediction system of the present invention, the test result acquisition unit in the test terminal is the subject's age, sex, educational history, caregiver, onset of dementia on the test date of the cognitive function test. Cognitive function may be affected by at least one selected from the group consisting of age, elapsed time from onset of dementia to taking antidementia, type and amount of antidementia taking, and past history Obtained as an influencing factor, the test result transmitting unit in the inspection terminal transmits including the influencing factor acquired by the test result acquiring unit, and the test result receiving unit in the inspection server receives including the influencing factor, The test result storage unit in the test server stores, for each subject, including the influence factor, and the prediction calculation formula estimation unit stores at least one of the stored influence factors. Select the subject who is the same as the elephant subject, and subject the above by non-linear regression analysis using the test score and the elapsed time from the onset to the test date in the past cognitive function test for the subject Estimate a prediction formula for predicting the future test score of the subject. With this form of system, it is possible to predict future cognitive functions using information unique to the subject.

  Although there is no strict limitation on the non-linear regression analysis method for estimating the prediction calculation formula in the cognitive function prediction system of the present invention, the prediction calculation formula estimation unit of the test server is stored in the test result storage unit. To predict future test scores of the above subjects by spline nonlinear regression analysis using test scores in the past cognitive function tests and the elapsed time from the onset to the test date for a large number of subjects The prediction data transmission unit of the test server transmits the test data including the test score in the past cognitive function test date for the target subject stored in the test result storage unit, The prediction data reception unit of the test terminal receives the test score including the test score on the past cognitive function test date for the target subject, and the prediction result generation unit of the test terminal uses the prediction calculation formula. The future test score of the target subject is generated by applying the prediction calculation formula by applying the test score on the past cognitive function test date and the current test score for the target subject. preferable. This nonlinear regression analysis technique enables prediction using the cognitive function transition of the subject subject to date.

  Dementia is defined as “a state in which an intellectual function that has been normally developed has continuously deteriorated due to an acquired brain disease, causing problems in daily life (social life)”. In the cognitive function prediction system, the future cognitive function prediction of the target subject is displayed on the test terminal in the form of prediction of the future test score, so the future status of the test subject is the user of the test terminal It may not be fully understood by healthcare professionals, subjects or caregivers. Therefore, in a preferred embodiment of the cognitive function prediction system of the present invention, the test result acquisition unit in the test terminal gives an alternative answer from the caregiver to a plurality of questions for evaluating the daily life function of the subject. Further, the inspection result transmission unit in the inspection terminal transmits the response including the answer acquired by the inspection result acquisition unit, and the inspection result reception unit in the inspection server receives the response including the answer. The test result storage unit stores the answer including the answer, and the test server uses the test score stored in the test result storage unit and the answer to obtain a correlation between the test score and the daily life function. The analysis unit and the correlation between the test score obtained by the correlation analysis unit and the daily life function and the response to the subject whose daily life function corresponding to the above question has deteriorated A correlation storage unit that stores the correlation, the prediction data transmission unit in the inspection server transmits the correlation and the comment stored in the correlation storage unit, and the inspection terminal A prediction data receiving unit including the correlation and the comment, and a prediction result generating unit in the inspection terminal refers to the correlation and the comment, and the future test score of the target subject A comment corresponding to the correlated daily life function is extracted, and the prediction result display unit in the inspection terminal displays the comment extracted by the prediction result generation unit as a comment for a predicted daily life function decline in the future.

  Multiple questions for assessing a subject's daily life functions include, for example, “Can I make big purchases or pay bills alone” or “Can I go to a new location alone?” ”Is shown to the subject's caregiver, and the caregiver's answer to such a question is acquired by either“ can ”or“ cannot ”.

  Then, the correlation analysis unit of the test server refers to the test score of the cognitive function test and the answer when the test score is acquired. For example, when the test score reaches a1, the customer pays a large purchase or pays the invoice. If the prediction result display unit in the test terminal finds that the future test score of the subject subject is a1, as a comment on the predicted daily life function decline in the future, Please note that it will be difficult to make large purchases or pay bills by yourself. " In addition to comments indicating the state of reduced daily functions, this comment may include advice such as precautions, training methods for maintaining daily life functions, and life rhythms recommended for maintaining daily life functions. .

  This display can prompt the preparation of the subject and the caregiver's mind, and medical staff can also give advice according to the state of daily living functions. In addition, identification of the onset time of dementia may be difficult for the subject or caregiver, but if an evaluation item that shows a decrease in daily life function is identified in the early stage of onset of dementia by the correlation analysis unit The onset time is easily identified by asking the caregiver a question corresponding to the item. In addition, even when the onset time has already been identified by interviewing the subject or caregiver, the question corresponding to the evaluation item in which the decline in daily life function is recognized at the beginning of the onset is again asked to the caregiver, If the obtained answer is different from that in the initial interview, the onset time data stored in the test result storage unit of the test server is corrected to the newly obtained data, and the onset time is specified. Since variation can be suppressed, as a result, the accuracy of prediction of future test scores of the subject subject can be improved.

  In the cognitive function prediction system of the present invention, the future test score of each subject is predicted, but not only the prediction of the future cognitive function of each subject but also the cognitive function in many people with dementia. If the trend of decline is understood, data that provides an objective grasp of the transition of cognitive function over time from the onset of dementia will be provided. It is preferable because it makes it possible to encourage the preparation of the minds of family members who are caring for and care, and also enables medical professionals to give appropriate advice.

  Therefore, in a preferred embodiment of the cognitive function prediction system of the present invention, the test server determines whether the test score for a large number of subjects stored in the test result storage unit and the elapsed time from the onset to the test date. The cognitive function transition analysis unit that calculates the relationship between the elapsed time from the onset and the test score by linear regression analysis, and the relationship between the elapsed time from the onset calculated by the cognitive function transition analysis unit and the test score An analysis result transmitting unit for transmitting to the analysis terminal, and the test terminal receives an analysis result receiving unit for receiving a relationship between an elapsed time from the onset transmitted from the test server and a test score, and receiving the analysis result And an analysis result display unit for displaying the relationship between the elapsed time from the onset received by the unit and the test score. The non-linear regression analysis in the cognitive function transition analysis unit may target all the data of the subjects stored in the test result storage unit of the test server, or may target the data of some subjects. The nonlinear regression analysis method in the cognitive function transition analysis unit is not strictly limited, and may be the same as or different from the nonlinear regression analysis method in the prediction arithmetic expression estimation unit. With this form of the cognitive function prediction system, data that objectively knows the transition of the cognitive function is visually provided. In addition, for example, the cognitive function transition analysis unit of the examination server is divided into a group of subjects who are taking treatment such as medication and a group of subjects who are not taking treatment, and the elapsed time from the onset and the examination score If the relationship is calculated, the effectiveness of the treatment can be visually confirmed, and the cognitive function transition analysis unit of the test server can determine the elapsed time from the onset and the test score for each group of subjects with different age at onset. If the relationship with is calculated, it is possible to visually confirm the effect of age at onset on cognitive function transition.

  The cognitive function test used in the cognitive function prediction system of the present invention is not particularly limited as long as it is a test that can evaluate the cognitive function, such as the cognitive function test shown in Table 1, and the system of the present invention is a server- Because it is a client system, it is easy to upgrade the version such as new cognitive function tests or new test items. However, because it is a system that predicts the transition of cognitive function, the course of dementia such as ADAS-Jcog It is preferred to use a test designed specifically for observation.

  In the cognitive function prediction system of the present invention, data such as the test score of each subject's cognitive function test acquired by the test result acquisition unit of the test terminal is directly transmitted to the test result acquisition unit by the user of the test terminal, for example. Although it may be input, it is preferable to perform a cognitive function test using a test terminal because a cognitive function test and data acquisition for a large number of subjects can be performed easily and quickly.

  Therefore, in a preferred embodiment of the cognitive function prediction system of the present invention, the test server scores test information including a method for evaluating the response of the subject in the cognitive function test and the response of the subject in the cognitive function test. A test content storage unit that stores score data for conversion into a test content, and a test content transmission unit that transmits test information and score data stored in the test content storage unit to the test terminal, The inspection terminal further includes an inspection content receiving unit that receives the inspection information and score data transmitted from the inspection server, and an inspection content display unit that displays the inspection information received by the inspection content receiving unit. The test result acquisition unit in the test terminal displays the response result of the subject in the cognitive function test performed according to the test information displayed on the test content display unit. It scored based on the number data received by the receiver, to obtain the rating result obtained as the test scores. A plurality of questions for evaluating the daily life function of the subject described above are also stored in the examination content storage unit, and questions are asked to the caregiver via the examination content transmission unit, the examination content reception unit, and the examination content display unit. Items may be provided, and the test result acquisition unit may acquire an alternative answer of the caregiver for this question.

  Cognitive function tests should be accurately performed by medical personnel such as doctors, clinical psychologists, and speech auditors, but there are surprisingly few medical workers skilled in cognitive function tests. There is a problem that the evaluation results vary greatly depending on the proficiency level of the workers. In order to solve this problem, the applicant is connected to a medical worker terminal used by a medical worker in Japanese Patent Application No. 2009-113223, which is unpublished at the time of this application, and this terminal via a communication line. A system for supporting dementia testing by health care workers, including a test support server. In this system, for each test item included in the test for dementia to be performed, prior knowledge for performing the test, precautions for performing the test, actions to be taken by the health care worker in performing the test, and subjects to be tested. The test information including the method for evaluating the examiner's response is displayed on the medical staff terminal from the test support server, and the medical staff only performs the dementia test according to the test information indicated, and the dementia test for the subject is performed. Result can be obtained. Therefore, even if it is a medical worker who does not have sufficient knowledge about the dementia test to be performed, or a medical worker who performs a dementia test for the first time, based on the indicated test information An accurate test can be easily performed, a uniform test result can be obtained no matter who performs the test, and the test result does not vary depending on the proficiency level of the medical staff.

  Also in the cognitive function prediction system of the present invention, the test information stored in the test content storage unit of the test server is prior knowledge for performing the test for each test item included in the cognitive function test, and the execution of the test The test results obtained by the test terminal are displayed on the test content display unit, including the precautions in the test, the behavior that the medical staff should perform in performing the test, and the method of evaluating the response of the subject in the test. Based on the score data received by the test content receiving unit, the response results of the test subjects included in the cognitive function test evaluated by the medical staff according to the information are scored, and the obtained score result is tested Even if it is a medical worker who does not have enough knowledge about the cognitive function test to be performed, Even a health care worker can easily perform an accurate test based on the indicated test information and can obtain a uniform test result no matter who performs the test. Because test results do not vary depending on the level of proficiency, test scores that are consistent for individual subjects can be obtained, and as a result, the accuracy of predicting future test scores for the target subjects is improved. Can be made.

  In the cognitive function prediction system of the present invention, the future test score of the target subject is predicted by utilizing the transition of the test score of the cognitive function test for a large number of subjects, so individual dementia patients The future cognitive function can be accurately predicted, and the future cognitive function can be predicted using information unique to the subject.

It is a figure which shows the outline of a structure of the cognitive function prediction system of embodiment of this invention. It is a functional block diagram of the cognitive function prediction system shown in FIG. It is a flowchart which shows the process which performs the cognitive function test | inspection about each subject, and daily life function evaluation. It is a display for acquiring basic information of a subject who takes a cognitive function test for the first time. This is a display for adding / updating basic information of each subject. This is a display that provides prior knowledge for the tests of "verbal language ability", "auditory understanding of language", and "difficulty in spoken language". This is a display that provides precautions for the tests of "verbal language ability", "auditory understanding of language", and "difficulty of speech in spontaneous speech". This is a display designating a question to be answered by a medical worker and an evaluation method of a subject's response in a test of “verbal language ability”, “auditory understanding of language”, and “difficulty speaking in spontaneous speech”. It is a figure which shows the question for a subject's daily life function evaluation, and the answer by a caregiver. It is a figure which shows notionally the data table which memorize | stores the test result of each subject. It is a flowchart which shows the process of obtaining the relationship between the test score in many subjects, and the elapsed time from onset. It is a display which shows the relationship between a test score and the elapsed time from onset. It is a flowchart which shows the process of performing prediction of the cognitive function and daily life function of a subject subject. It is a figure which shows notionally the data table which memorize | stores the test score correlated with the comment corresponding to the question for evaluating the subject's daily life function, and the daily life function corresponding to a question. It is a display which shows the prediction of the future cognitive function about a subject subject. It is a display which shows the comment with respect to the daily life function fall estimated in the future with the prediction of cognitive function.

  The cognitive function prediction system of the present invention is connected to a test terminal for a test of a subject's cognitive function, and this test terminal via a communication line, for predicting the future cognitive function of each subject. And an inspection server that provides various information to the inspection terminal. Hereinafter, an embodiment of the present invention will be described by way of an example in which a medical staff uses a test terminal, ADAS-Jcog is used as a cognitive function test, and daily life functions are simultaneously evaluated. The present invention is not limited to this form, and modifications can be made without departing from the spirit of the present invention.

  First, ADAS-Jcog corresponding to the Japanese version of ADAS-cog will be described. ADAS-cog, which is the basis, is a cognitive function test scale that has been proposed for the purpose of evaluating changes in cognitive function by cholinergic drugs in individuals with Alzheimer-type dementia (AD). ”,“ Auditory comprehension of language ”,“ difficulty speaking in spontaneous speech ”,“ word reproduction ”,“ following verbal instructions ”,“ hand and article designation ”,“ constructive action ”,“ ideal movement ”,“ orientation ” , “Word recognition” and “test teaching reproduction ability” are included. ADAS-Jcog is almost equivalent to ADAS-cog, but is frequently used in the oral language of Japanese elderly people as words in “word reproduction”, “word recognition”, and “(hand and) article designation”. This test scale is improved so that Japanese can be evaluated more appropriately by adopting words.

  The “verbal language ability” test is a test that evaluates the qualitative aspects of a subject's speech, such as the clarity of words and the ability to understand others, and it is an unclear speech in the overall conversation. Evaluate the percentage of words or unknown utterances. The “auditory understanding of language” test is a test for evaluating the ability of the subject to understand the words spoken by the medical staff, and evaluates the frequency of comprehension disorder. The “spoken difficulty in self-speaking” test is a test that evaluates whether words to be said in a subject's self-speech appear, and evaluates the frequency of speech-speaking disorders.

  The “word reproduction” test is a medical professional who presents and reads 10 words written on a card to a subject one by one, and then remembers the words read by the subject. The same test is repeated three times, and the average number of incorrect answers is evaluated. “Follow verbal instructions” test means that a health care worker “holds the fist”, “points to the ceiling, then points to the floor”, “with both eyes closed with two fingers “Strike your shoulders twice”, “Place the pencil on the white paper, then put it back”, “Place the watch on the other side of the pencil and turn the white paper over” This is a test that evaluates the auditory comprehension of oral language through the ability of the subject to instruct verbal actions in order and the subject's ability to execute it, and evaluates the number of instructions that the subject could not execute. .

  The “hand and article designation” test is a test that evaluates the number of incorrect answers by asking the subject the names of the five fingers of the subject's dominant hand and the names of twelve articles shown to the subject at random. is there. Twelve articles, four high-frequency articles (chairs, bicycles, scissors, kanachichi), four medium-frequency articles (claws, combs, abacus, brush), and four, depending on the frequency of appearance in daily life Of low-frequency articles (handkerchiefs, notebooks, rings, fans). The “constructive action” test is a test for evaluating the ability of the subject to replicate a graphic composed of a circle, two overlapping rectangles, a diamond, and a cube, and evaluates the number of figures that could not be copied. The “ideal movement” test is a test of giving a letter to a subject, an envelope, and a stamp, and assuming that a letter will be issued, folding the letter, putting the letter in the envelope, sealing the envelope, and assigning the name to the envelope It is a test that evaluates whether you can write and put stamps on envelopes, and evaluates the number of actions that could not be done.

  The “registration” test is a test that asks the subject about eight items of year, month, day, day of the week, time, season, place, and person and evaluates the number of incorrect answers. The “word recognition” test is a 12-card card with specific words written by the health care professional, presenting them to the subject and reading them aloud. A total of 24 cards, including 12 cards with new words written on them, are randomly presented to the subject one by one, allowing the subject to identify whether or not it is the first presented word, and This is a test for evaluating the number of correct answers, and the same test is repeated three times to evaluate the average number of incorrect answers. The “test teaching regeneration capability” is a test for evaluating whether or not the subject remembers the teaching of the medical staff to the subject in the word recognition test, and evaluates the number of times the teaching is forgotten.

  ADAS-Jcog scores do not reflect age and years of education, and can assess AD severity regardless of these, and further, ADAS-Jcog scores on repeated ADAS-Jcog scores This is a very useful test because the degree of progression and the effect of treatment can be confirmed.

  A plurality of questions for evaluating the daily life function of a subject include items that are expected to be correlated with cognitive decline in ICF (International Classification of Functioning, Disability and Health). It can be extracted and created, and the existing daily life movement evaluation scale, for example, Instrumental Activities of Daily Living Scale (IADL), Disability Assessment for Dimension for Sentiment (DAD), and Alzheimer's Measure ADL), Physical Self-Maintenan It is possible to extract and create items that are expected to correlate with cognitive function from se Scale (PSMS) and the N-style elderly life activity rating scale (N-ADL). Examples of questions for daily life function evaluation include "Can I make big purchases or pay my bill alone", "Can I go to a new place alone", " "Is the promise properly?" "Can I do my daily shopping alone?" "Can I use the cooler remote control?" "Can I operate the TV remote control?" ”,“ Can you make a simple menu ”,“ Can you eat without assistance ”,“ Do you sometimes miss the toilet? ”,“ You can choose your own clothes and change clothes "Can you manage your daily money yourself?", "Can you talk freely?"

  FIG. 1 shows an outline of the configuration of the cognitive function prediction system of the present embodiment. The cognitive function prediction system 1 of the present invention includes an inspection server 10 and many inspection terminals 30 having a touch panel display, and the inspection server 10 and the inspection terminals 30 are connected by a communication line 50 such as the Internet. . The inspection terminal 30 is arranged in a hospital or the like, and a medical worker performs a cognitive function inspection while looking at the terminal 30. The present embodiment is a server-client system, and the inspection terminal 30 operates as a terminal having the following functions by application software provided from the inspection server 10 to the inspection terminal 30.

  FIG. 2 is a functional block diagram of the cognitive function prediction system 1 including the inspection server 10, the inspection terminal 30, and the communication line 50 shown in FIG. For ease of understanding, only one inspection terminal 30 is shown.

  The test server 10 constituting the cognitive function prediction system 1 includes a test content storage unit 11, a test content transmission unit 12, a test result reception unit 13, a test result storage unit 14, a cognitive function transition analysis unit 15, an analysis result transmission unit 16, The prediction calculation formula estimation unit 17, the prediction calculation formula storage unit 18, the correlation analysis unit 19, the correlation storage unit 20, and the prediction data transmission unit 21 are configured.

  The examination content storage unit 11 should evaluate the prior knowledge for performing the test for each of the 11 test items included in the ADAS-Jcog described above, that is, what the subject is doing in this test. Specific contents such as what should be evaluated and what should not be evaluated; Precautions in performing the test, that is, specific instructions for preventing erroneous test execution; Specific actions to be taken by the health care professional, such as what to ask the subject in the test, what to tell the subject, what to show to the subject; and Evaluation method of test subject's response, that is, “○” is input as the evaluation result if the answer to the question from the medical staff is correct, and “×” is input as the evaluation result if the answer is incorrect What specific designation content is stored as examination information, and further, score data for converting a subject's response in ADAS-Jcog into scores, and the above-described daily life function of the subject A plurality of questions are memorized.

  The inspection content transmitting unit 12 is a prior knowledge for performing a test and performing a test regarding each test of 11 test items included in ADAS-Jcog stored as inspection information in the inspection content storage unit 11. Notes on medical care, behaviors that medical staff should make in conducting tests, test information consisting of methods for evaluating test subject responses, and scores for converting subject responses in ADAS-Jcog into points The data is transmitted to the examination terminal 30 together with a plurality of questions for evaluating the daily life function of the subject described above.

  The test result receiving unit 13 is a test score obtained by scoring according to the score data (ADAS-) the evaluation result for the response of the test subject by the medical worker obtained in the process in which the medical worker performed ADAS-Jcog according to the test information. Jcog's total score), the day (examination date) when this examination score was obtained, and the caregiver's answers to the above-mentioned multiple questions for evaluating the subject's daily life function from the examination terminal 30 To receive. The test result receiving unit 13 further includes, as basic information unique to each subject, an identification number for identifying the subject, name, age, sex, educational history, presence of a caregiver, and if present Caregiver, timing of onset of dementia, presence or absence of medication, type and amount of antidementia drug being taken if medication is allowed, elapsed time from onset of dementia to taking antidementia, history of history The presence / absence of cerebrovascular disorder and the presence / absence of mental illness are received from the test terminal 30. The test result storage unit 14 includes, for each subject, a plurality of information for evaluating the basic information of the subject, the test score on the cognitive function test date, and the daily life function received by the test result receiving unit 13. The answer to the question is stored in correspondence with the inspection date.

  The cognitive function transition analysis unit 15 analyzes the transition of cognitive function in the population of people with onset of dementia, that is, the relationship between the elapsed time from the onset and the test score, and a large number stored in the test result storage unit 14. Data on the subjects of the subject, the onset time of the dementia of the subject, the test score of the cognitive function test date, and the influencing factors on the cognitive function included in the basic information (eg, age at onset, treatment from onset Elapsed time) is calculated, the elapsed time from the onset (= examination date-onset timing) is calculated, and the relationship between the elapsed time from the onset and the examination score is estimated by spline regression and stored. The analysis result transmission unit 16 reads the relationship between the elapsed time from the onset and the test score estimated by the cognitive function transition analysis unit 15 and transmits this to the test terminal 30. .

  The prediction calculation formula estimation unit 17 is included in the dementia onset time, the test score of the past cognitive function test date, and the basic information among the data about a large number of subjects stored in the test result storage unit 14. Factors affecting cognitive function are read out, the elapsed time from the onset (= examination date-onset time) is calculated, and nonlinear regression analysis using a spline regression model is performed to estimate the prediction formula. The prediction calculation formula storage unit 18 stores the prediction calculation formula estimated by the prediction calculation formula estimation unit 17.

  The correlation analysis unit 19 includes a plurality of test scores and data for evaluating the daily life function on the test day on which the test score is obtained, among the data about a large number of subjects stored in the test result storage unit 14. The answer to the question is read out, and the correlation between the test score and the daily life function is obtained. The correlation storage unit 20 stores the correlation between the test score obtained by the correlation analysis unit 19 and the daily life function together with comments on the subject whose daily life function corresponding to the above question has deteriorated. . The prediction data transmission unit 21 calculates the test score of the past cognitive function test date for the target subject that predicts the future cognitive function among the data on the large number of subjects stored in the test result storage unit 14. The read data is transmitted to the test terminal 30 together with the prediction calculation formula read from the prediction calculation formula storage unit 18, the correlation between the test score read from the correlation storage unit 20 and the daily life function, and the above comment. It is.

  The inspection terminal 30 includes an inspection content reception unit 31, an inspection content display unit 32, an inspection result acquisition unit 33, an inspection result transmission unit 34, an analysis result reception unit 35, an analysis result display unit 36, a prediction data reception unit 37, and a prediction result generation. The unit 38 and the prediction result display unit 39 are included.

  The examination content reception unit 31 includes a plurality of questions transmitted from the examination content transmission unit 12 of the examination server 10 for evaluating the daily life function of the subject, and 11 test items included in the ADAS-Jcog. For each test, test information consisting of prior knowledge to perform the test, precautions in performing the test, actions to be taken by medical personnel in performing the test, and methods for evaluating the response of the subject in the test Is received together with score data for converting the response of the subject in ADAS-Jcog into a score.

  The examination content display unit 32 is a prior knowledge for performing a test, notes on performing the test, a test regarding each test of the 11 test items included in the ADAS-Jcog received by the examination content receiving unit 32 In order to evaluate the daily life function of the subject and display the test information consisting of the behavior and behavior of the medical staff in the implementation of the test and the method of evaluating the response of the subject in the test sequentially as the test progresses Display multiple questions.

  When the test result acquisition unit 33 has basic information unique to each subject, an identification number for identifying the subject, name, age, sex, educational history, presence or absence of a caregiver, or caregiver Of the caregiver, timing of onset of AD, presence or absence of medication, type and amount of antidementia drug being taken if medication is allowed, elapsed time from onset of AD to taking antidementia, history of history Acquire the presence / absence, presence / absence of cerebrovascular disorder, presence / absence of mental illness, and accept evaluation results on the response of the subject by the medical staff obtained in the process of the medical staff conducting the test according to the above test information In addition, the test content receiving unit 31 scores the test subject's response according to the score data for converting the score into scores, and obtains the scoring result (ADAS-Jcog total score) as the test score along with the test date. The daily life function of the examiner Worthy caregivers "can" with respect to a plurality of questions for, is to get the alternatively selected answer from the choices of "can not". The test result transmitting unit 34 sends, to the test server 10, answers to a plurality of questions for evaluating the basic information of the subject, the test date and the test score, and daily life functions acquired by the test result acquiring unit 33. To be sent.

  The analysis result receiving unit 35 receives the relationship between the elapsed time from the onset and the test score analyzed by the cognitive function transition analysis unit 15 of the test server 10 from the test server 10, and the analysis result display unit 36 receives The relationship between the elapsed time from the onset and the test score is displayed. The prediction data receiving unit 37 includes a prediction calculation formula estimated by the prediction calculation formula estimation unit 17, a test score of a past cognitive function test date for a subject who predicts a future cognitive function, and a correlation analysis unit. The test score obtained by 19 and the correlation between the daily life function and the comment on the subject whose daily life function has deteriorated are received from the inspection server 30.

  The prediction result generation unit 38 adds the test score in the past cognitive function test date received by the prediction data receiving unit 37 and the test result acquiring unit 33 to the prediction calculation formula received by the prediction data receiving unit 37. The test score received by the prediction data receiving unit 37 is generated by applying the current test score and the influencing factor of the target subject acquired by the prediction data and generating the prediction and error range of the test score of the future target subject. Refer to comments for subjects with reduced daily life functions and daily life functions, and comments corresponding to daily life functions that correlate with future test scores of the subject subject. It is extracted as a comment on the decline in living functions. The prediction result display unit 39 predicts the future test score and the error range for the subject subject generated by the prediction result generation unit 38, and the daily life predicted in the future similarly extracted by the prediction result generation unit 38. It is displayed with a comment on the decline in life function.

  Next, specific processing in the cognitive function prediction system 1 of the present embodiment will be described.

(1) Acquisition of basic information, test result of cognitive function test, and evaluation result of daily life function Figure 3 shows basic information about each subject, test score of cognitive function test on test day and daily life function It is the flowchart which showed the process of acquiring the evaluation result and.

  When a medical worker who plans to implement ADAS-Jcog activates the test terminal 30 of the cognitive function prediction system 1 and accesses the test server 10 via the communication line 50 (S1), the test of the test server 10 A plurality of questions for evaluating deterioration of the daily life function of the subject stored in the content storage unit 11 and tests for each of the 11 test items of ADAS-Jcog. Test information consisting of prior knowledge, precautions for test execution, behaviors that medical personnel should perform in test execution, and methods for evaluating the response of subjects in the test, scores the responses of the subjects in ADAS-Jcog Along with the point data for conversion into the inspection content transmission unit 12 of the inspection server 10 is transmitted (S2) to the inspection content reception unit 31 of the inspection end 30 Is Shin (S3).

  In the cognitive function prediction system 1 according to the present embodiment, first, basic information of a subject is acquired via a touch panel display that functions as the test result acquisition unit 33. FIG. 4 shows a screen display of a touch panel display for acquiring basic information of a subject who takes an ADAS-Jcog test for the first time. As basic information, identification number to identify the subject, name, age, gender, educational history, presence or absence of caregiver, if there is a caregiver, the caregiver, AD onset time, presence or absence of medication, medication If the drug is found, the date of start of medication and the type and amount of the antidementia drug being taken, the presence / absence of a history, presence / absence of cerebrovascular disorder, and presence / absence of mental illness are obtained. Age and years of education are automatically calculated. Among the basic information, information on gender, age, AD onset time, educational history, caregiver, medication start date, type and amount of nootropics, history of presence, presence of cerebrovascular disorder, presence of mental illness It is acquired as an influencing factor that can affect cognitive function. The medical record number is treated as an identification number for this subject. When the AD onset time is uncertain, data is input in a form such as “around summer” and “around winter”. As for cerebrovascular disorders and psychiatric disorders, typical examples of diseases are specified, and the presence or absence of diseases can be accurately input. FIG. 5 shows a screen for adding / updating basic information of a subject who has taken ADAS-Jcog in the past. Of the basic information shown in FIG. 4, only basic information that can change over time is input. The descriptions in the left and right frames in FIGS. 4 and 5 show options that can be selected in each item for convenience. The medical staff listens to the basic information from the subject or caregiver while viewing the screen, and inputs the interview results in the corresponding part.

  In the cognitive function prediction system 1 according to the present embodiment, the cognitive function test using ADAS-Jcog is performed prior to the evaluation of daily life function, and the test information received by the test content receiving unit 31 is the test content display unit. It is displayed sequentially on the touch panel display that also functions as 32 as the examination progresses.

  In ADAS-Jcog, first, tests of three items of “verbal language ability”, “auditory understanding of language” and “difficulty speaking in spontaneous speech” are performed at the same time. It is evaluated through free conversation with the subject. Because of the free conversations in these tests, questions related to “physical condition”, “mood”, “audiovisual”, “handedness”, “hobby”, “life history”, “current life”, “examination day” 18 are prepared, and 5 to 8 of the 18 questions are evaluated for the subject's "verbal language ability", "auditory understanding of language", and "difficulty speaking in spontaneous speech". Is done.

  FIGS. 6 to 8 show the test content display unit 32 and the test result acquisition unit of the test terminal 30 for performing the tests of “verbal language ability”, “auditory understanding of language”, and “difficulty speaking in spontaneous speech”. The screen display of the touch panel type display which functions as 33 is shown. The upper part of FIG. 6 is a display of prior knowledge (S5), the upper part of FIG. 7 is a display of precautions (S6), and the upper part of FIG. 8 is a question from the medical staff to the subject in these tests. The question is to be displayed (S7), and the lower part is a display that combines the display of the evaluation method (S8) and the input part of the evaluation result by the medical staff.

  In the upper part of the screen display example in FIG. 8, a question “How are you feeling?” And a question “Are there any painful places?” Are shown, “How are you feeling?” The date of “2008/10/21” is shown before the display. This display indicates that a question “How are you feeling?” Was asked in a past examination conducted on the same subject on October 21, 2008. In the "verbal language ability", "auditory understanding of language", and "difficulty speaking in self-speaking" tests, the same question should be selected in advance because the same question should in principle be repeated for the same subject. Has been displayed. This selection is performed with reference to the identification number of the subject input on the basic information acquisition screen of FIGS. 4 and 5 and the past test results stored in the test result storage unit 14 of the test server 10. .

  While looking at the screen display of this touch panel display, the health care professional gained prior knowledge for the tests of "verbal language ability", "auditory understanding of language" and "difficulty in spoken language" Confirm and conduct the test by asking the questions shown on the display. For example, even if a medical staff asks a question, if the subject is silent, the medical staff touches the “silent” display part (see the lower part of FIG. 8) of the touch panel display, and the evaluation result of “silent” is It is received by the inspection result acquisition unit 33 (S9). Next, "Word Replay", "Follow verbal instructions", "Hand and article designation", "Constitutional action", "Ideal movement", "Registration", "Word recognition", "Test teaching reproducibility" Is performed in the same manner, and after a series of tests are completed (S10), the evaluation result in each test is scored with reference to the score data received by the examination content receiving unit 31, and the total of the scores of each test is totaled The score is acquired as an inspection score (S11).

  In the cognitive function prediction system 1 of the present embodiment, the daily life function of the subject is evaluated after the cognitive function test by ADAS-Jcog. FIG. 9 shows a screen display of a touch panel display that functions as the examination content display unit 32 and the examination result acquisition unit 33 of the examination terminal 30 for evaluating the daily life function of the subject. While looking at the screen, the health care worker asks the questions displayed on the screen, such as “Can I do everyday shopping alone” or “Can I operate the cooler remote control”? The answer selected from “Can” and “Can't” with respect to these questions is input to the corresponding part (S12).

  The ADAS-Jcog test score of the subject obtained in the above process and the answer to the question for evaluating the daily life function are transmitted from the test result transmission unit 34 of the test terminal 30 (S13), and the test server 10 Is received by the test result receiving unit 13 (S14), and stored in the test result storage unit 14 for each subject (S15). FIG. 10 shows a data table for each subject stored in the test result storage unit 14. Basic information that can change depending on the passage of time in addition to identification number, name, gender, date of birth, AD onset, and educational history as basic information that does not depend on passage of time for each subject Presence / absence of medication start / change, type and amount of drugs used, caregiver, presence / absence of medical history, presence / absence of cerebrovascular disorder, presence / absence of mental illness, ADAS-Jcog test scores, and daily life function The answer to each question is stored in correspondence with the inspection date. When the AD onset time is uncertain and data is acquired in the form of “around summer” or “around winter”, the intermediate month in each season is stored in the data table.

(2) Cognitive Function Transition Analysis FIG. 11 is a flowchart showing a process for obtaining the relationship between the test score and the elapsed time from onset in many subjects. In the following, we will explain the example of investigating the influence of the elapsed time from the onset of dementia to taking an antidementia drug on the relationship between the test score and the elapsed time from the onset, but the data subject to nonlinear regression analysis Various changes to the relationship between the test score and the elapsed time since onset by changing the group of subjects to be provided, such as age, sex, subjects with and without cerebrovascular disorder The influence of various factors can be investigated.

  The cognitive function transition analysis unit 15 of the test server 10 reads data on AD onset time and medication start time from the data of many subjects stored in the test result storage unit 14 (S101). Calculate the elapsed time until taking the dementia drug (= starting time of medication-onset time), and many subjects were treated with the group of subjects who started medication 3 years after the onset, and 5 years after the onset of medication. It is divided into a group of subjects who started and a group of subjects who started medication 10 years after the onset. The cognitive function transition analysis unit 15 further reads the test date and the test score in the past cognitive function test for the subjects belonging to each group (S101), and the elapsed time from the onset to the test date (= test date-onset time) ) Is calculated (102), nonlinear regression analysis is performed, and the relationship between the elapsed time from the onset and the examination score is analyzed (S103), and the obtained result is stored (S104).

  In the cognitive function transition system 1 of the present embodiment, the non-linear regression analysis by the cognitive function transition analysis unit 15 is performed by spline regression.

をあてはめる場合、ｆ（ｘ）として以下の式で表される２次スプライン関数を仮定する。

ここで、Ｚは、予め定められた影響因子（説明変数）を含むベクトルであり、βは、未知の回帰係数である。ｓ_１，・・・・，ｓ_ｒはスプラインの節点であり、（ｘ−ｓ_ｋ）_＋は、ｘ＞ｓ_ｋならばｘ−ｓ_ｋ、ｘ≦ｓ_ｋならば０を返す関数である。そして、ｆ（ｘ）は、以下の損失関数を最小にするような回帰係数δ_０，δ_１，δ_２，ａ_１，・・，ａ_ｒ，βを既存の最適化ルーチンを用いて探索することにより得られる。

ここで、λは平滑化パラメータである。誤差範囲は、既存の手法を用いてｆ（ｘ）＋βＺの９５％信頼区間を計算することにより得られる。 In the cognitive function transition system 1 of the present embodiment, a set of observations (x _i , y _i ) (i = 1,..., Elapse time from the onset (x) and examination score (y). n) for nonlinear regression models

Is applied, a quadratic spline function represented by the following equation is assumed as f (x).

Here, Z is a vector including a predetermined influence factor (explanatory variable), and β is an unknown regression coefficient. s ₁ ,..., s _r are spline nodes, and (x−s _k ) ₊ is a function that returns x−s _k if x> s _{k and} 0 if x ≦ s _k . Then, f (x) searches for regression coefficients δ ₀ , δ ₁ , δ ₂ , a ₁ ,..., A _r , β that minimize the following loss function using an existing optimization routine. Can be obtained.

Here, λ is a smoothing parameter. The error range is obtained by calculating a 95% confidence interval of f (x) + βZ using an existing method.

  The cognitive function transition curve obtained as a result of the non-linear regression analysis by the cognitive function transition analysis unit 15 is sent to the analysis result transmission unit 16 and transmitted from the analysis result transmission unit 16 to the test terminal 30 (S105). The cognitive function transition curve received by the analysis result receiving unit 35 of the terminal 30 is displayed on the screen of the touch panel display that also functions as the analysis result display unit 36 (S107). FIG. 12 is a screen display of the touch panel type display showing the obtained cognitive function transition curve. The group of subjects who started the medication 3 years after the onset, the subjects who started the medication 5 years after the onset The cognitive function transition curves obtained for each of the group of examiners and the group of subjects who started administration 10 years after onset are shown.

(3) Prediction of Future Cognitive Function and Daily Life Function FIG. 13 is a flowchart showing a process for predicting the cognitive function and daily life function of the subject subject. This process is performed immediately after the subject's current cognitive function test is performed.

  The prediction calculation formula estimation unit 17 of the test server 10 includes AD onset time, influencing factors, and test scores on the past cognitive function test dates from the data of many subjects stored in the test result storage unit 14. Data is read out (S201), the elapsed time from the onset to the examination date (= examination date-onset time) is calculated (S202), the spline regression analysis including the influence factors of each subject is performed, and the prediction formula Is estimated (S203).

ここで、Ｙは対象被検者の将来の検査得点であり、Ｘ_１，・・・，Ｘ_Ｋは今回及び過去Ｋ−１回の検査日（発症時期から（検査日−発症時期）だけ経過した時点）の検査得点であり、ｔは発症からの経過時間とする。Ｚは、予め定められた影響因子（説明変数）を含むベクトルである。βは、被検者の過去の認知機能検査日の検査得点と影響因子とから推定される未知の回帰係数である。ｆ_ｋ（ｘ）（ｋ＝１，・・・，Ｋ＋１）は、被検者の過去の認知機能検査日の検査得点と影響因子とから推定される未知の関数である。
ｆ_ｋ（ｘ）としては以下に示す２次スプライン関数が用いられる。

ここで、κ_ｋ０,・・・・，κ_kＬは合計Ｌ＋１箇所のスプライン関数の節点である。α_ｋｏ,・・・，α_ｋＬは、被検者についての過去の認知機能検査日の検査得点と影響因子から推定される未知の回帰係数である。（ｘ−κ_ｋｌ）_＋は、ｘ＞κ_ｋｌならばｘ−κ_ｋｌ、ｘ≦κ_ｋｌならば０を返す関数である。 In the spline regression for estimating the prediction calculation formula in the cognitive function prediction system 1 of the present embodiment, the prediction calculation formula g (X ₁ ,..., X _K , t, Z) is expressed in the following linear format. Is done.

Here, Y is the future test score of the subject subject, and X ₁ ,..., X _K are the current and past K-1 test dates (examination date-(test date-onset time)) The t is the elapsed time from the onset. Z is a vector including a predetermined influence factor (explanatory variable). β is an unknown regression coefficient estimated from the test score and the influencing factor of the subject's past cognitive function test date. f _k (x) (k = 1,..., K + 1) is an unknown function estimated from the test score and the influence factor of the subject's past cognitive function test date.
The following quadratic spline function is used as f _k (x).

Here, κ _k0 ,..., Κ _kL are nodes of a total of L + 1 locations of the spline function. α _ko ,..., α _kL are unknown regression coefficients estimated from the test score and the influencing factors on the past cognitive function test date for the subject. (X-κ _{kl) +} is a function that returns x> kappa _kl if x-kappa _kl, the x ≦ kappa _kl if 0.

ここで、λは平滑化パラメータである。誤差範囲は、既存の手法を用いてｇ（Ｘ_１，・・・，Ｘ_Ｋ，ｔ，Ｚ）の９５％信頼区間又は９５％予測区間を計算することにより得られる。 In spline regression, the prediction formula is estimated by the following method from the test score and the influencing factors of the subject's past cognitive function test date. Y _ij is the score of the cognitive test at time point j (time point j after onset) of subject i, and X _ij1 ,..., X _ijK are the past K times of the cognitive test at time point j of subject i. Score. Let tij be the elapsed time from the onset of the time point j of the subject i. Let Z _{ij be} a vector containing the influencing factors of the time point j of the subject i. The prediction formula is obtained by searching for α _k0 ,..., Α _kL , β that minimizes the following loss function using an existing optimization routine.

Here, λ is a smoothing parameter. The error range is obtained by calculating a 95% confidence interval or a 95% prediction interval of g (X ₁ ,..., X _K , t, Z) using an existing method.

  The prediction calculation formula estimated by the prediction calculation formula estimation unit 17 is stored in the prediction calculation formula storage unit 18 (S204).

  On the other hand, the correlation analysis unit 19 of the test server 10 is used to evaluate the test score and the daily life function on the past cognitive function test date from the data of many subjects stored in the test result storage unit 14. An answer to the question is read, and a correlation between the examination score and the daily life function is obtained (S205). Specifically, the number of answers that answered “cannot be” for each question when a certain test score is obtained is calculated. Then, for example, in response to the question “Is it possible to pay alone for big purchases or bills” in question 1 shown in FIG. If the number increases rapidly, the test score a1 is associated with the daily life function corresponding to question 1, and the test score is given to the question “Can you go to a new place alone” in question 2 If the number of responses that “cannot be” is abruptly increased after a2 points, the test score a2 is associated with the daily life function corresponding to the question 2.

  The correlation between the daily life function and the examination score obtained by the correlation analysis unit 19 is stored in the correlation storage unit 20 having a data table in which comments corresponding to each question are stored in advance (S206). FIG. 14 conceptually shows a data table storing the question numbers shown in FIG. 9, comments corresponding to each question, and examination scores associated with daily life functions corresponding to each question. FIG.

  The prediction calculation formula stored in the prediction calculation formula storage unit 18, the correlation between the test score and the daily life function stored in the correlation storage unit 20, and the comment are sent from the prediction data transmission unit 21 to the test terminal 30. Thus, it is transmitted together with the test score of the subject subject's past cognitive function test date read from the test result storage unit 14 (S207), and these data are received by the prediction data receiving unit 37 of the test terminal 30 (S208). ).

に、受信した対象被検者の過去の認知機能検査日における検査得点と、検査結果取得部３３により取得された現在の検査得点及び影響因子を含むベクトルとを当てはめ、予測結果を生成する。 Based on the data received by the prediction data reception unit 37, the prediction result generation unit 38 of the inspection terminal 30 first generates the number of inspection points and error range predicted for the subject subject in the future (S209). Specifically, the prediction formula

Then, the received test score of the target subject on the past cognitive function test date and the current test score acquired by the test result acquisition unit 33 and the vector including the influencing factors are applied to generate a prediction result.

  The prediction result generation unit 38 also extracts a comment corresponding to the daily life function correlated with the predicted examination score based on the data received by the prediction data receiving unit 37 (S209). For example, if the current test score in the subject is a and a <a9 <a10 <a11 (in ADAS-Jcog, the cognitive function decreases as the score increases), the test score a9 is “daily. Extract the comment "Please note that it will be difficult to manage the money in general", and extract the comment "Please keep in mind that the frequency of not meeting the toilet will increase" for the test score a10, A comment is extracted with respect to the inspection score a11 that “the device is difficult because the conversation is difficult” (see FIG. 14).

  The generation result in the prediction result generation unit 38 is displayed on the screen of the touch panel display that also functions as the prediction result display unit 39 (S210). FIG. 15 shows a display screen that displays the results of calculating the prediction score and error range in the future cognitive function test of the subject subject with the presence or absence of medication as an influencing factor. FIG. The display screen displayed including a comment on the predicted decline in daily life function is shown.

  Although one embodiment of the present invention has been described above, the cognitive function prediction system of the present invention may use not only ADAS-Jcog but also the cognitive function test shown in Table 1, and the total score of the cognitive function test The score for each test item included in each cognitive function test may be used as the test score. Further, the nonlinear regression analysis method is not limited to that of the above-described embodiment. For example, a cognitive function transition curve (prediction formula) representing the relationship between the elapsed time from the onset and the test score as illustrated in FIG. 12 is obtained, and the current test score of the subject subject is represented in this cognitive function transition curve. When using a cognitive function test where the test score increases as cognitive function decreases, if the current test score is a, the part of the cognitive function transition curve where the test score is greater than a It can also be displayed as a prediction of the future cognitive function of the subject subject.

  The cognitive function prediction system of the present invention can be suitably used for predicting the future cognitive function of each individual with dementia.

DESCRIPTION OF SYMBOLS 1 Cognitive function prediction system 10 Test | inspection server 11 Test content memory | storage part 12 Test content transmission part 13 Test result reception part 14 Test result memory | storage part 15 Cognitive function transition analysis part 16 Analysis result transmission part 17 Prediction arithmetic expression estimation part 18 Prediction arithmetic expression memory | storage Unit 19 Correlation analysis unit 20 Correlation storage unit 21 Prediction data transmission unit 30 Inspection terminal 31 Inspection content reception unit 32 Inspection content display unit 33 Inspection result acquisition unit 34 Inspection result transmission unit 35 Analysis result reception unit 36 Analysis result display unit 37 Prediction data reception unit 38 Prediction result generation unit 39 Prediction result display unit 50 Communication line

Claims (8)

Including an inspection server and an inspection terminal connected via a communication line;
The test terminal is acquired by a test result acquisition unit that acquires a dementia onset time and a test date and a test score in a cognitive function test for a subject developing dementia, and the test result acquisition unit. A test result transmission unit that transmits the test subject's dementia onset time and the test date and test score in the cognitive function test to the test server;
The test server receives the test subject's dementia onset time and the test date and test score in the cognitive function test transmitted from the test terminal, and the onset of dementia for a large number of subjects. A test result storage unit storing the test date and test score in the cognitive function test for each subject,
It is a cognitive function prediction system for predicting the future test score of the subject subject whose current test score has been acquired by the test result acquisition unit of the test terminal,
The inspection server is
By non-linear regression analysis using the test score and the elapsed time from the onset to the test date in the past cognitive function test for a large number of subjects stored in the test result storage unit, A prediction formula estimator that estimates a prediction formula for predicting future test scores;
A prediction calculation formula storage unit that stores the prediction calculation formula estimated by the prediction calculation formula estimation unit;
A prediction data transmission unit that transmits the prediction calculation formula stored in the prediction calculation formula storage unit to the inspection terminal;
Further comprising
The inspection terminal is
A prediction data receiving unit for receiving a prediction calculation expression transmitted from the inspection server;
A prediction result generating unit that generates a future test score of the target subject by applying the current test score of the target subject to the prediction formula received by the prediction data receiving unit;
A prediction result display unit that displays a future examination score of the target subject generated by the prediction result generation unit;
Further equipped with,
Cognitive function prediction system characterized by this. The test result acquisition unit in the test terminal is the test subject's age, sex, educational history, caregiver, age at onset of dementia, elapsed time from onset of dementia to taking dementia drug Acquiring at least one selected from the group consisting of the type and amount of an antidementia drug being taken and a history of cerebrovascular disorder or mental illness as an influencing factor that may affect cognitive function,
The inspection result transmission unit in the inspection terminal transmits including the influence factor acquired by the inspection result acquisition unit,
The inspection result receiving unit in the inspection server receives including the influencing factors,
The test result storage unit in the test server stores for each subject including the influencing factors,
The prediction arithmetic expression estimating unit selects a subject in which at least one of the stored influencing factors is the same as the target subject, and a test score in a past cognitive function test for the selected subject The cognitive function prediction according to claim 1, wherein a prediction arithmetic expression for predicting a future test score of the target subject is estimated by a non-linear regression analysis using an elapsed time from an onset to an examination date. system. The spline using the test score and the elapsed time from the onset to the test date in the past cognitive function test for a large number of subjects stored in the test result storage unit by the prediction calculation formula estimation unit of the test server By non-linear regression analysis, a prediction formula for predicting the future test score of the target subject is estimated,
The prediction data transmission unit of the inspection server transmits including the test score on the past cognitive function test date for the target subject stored in the test result storage unit,
The prediction data receiving unit of the test terminal receives the test score including the test score in the past cognitive function test date for the subject subject,
The prediction result generation unit of the test terminal applies the test score and the current test score on the past cognitive function test date for the target subject to the prediction calculation formula, and executes the prediction calculation formula, The cognitive function prediction system according to claim 1, wherein a future test score of the target subject is generated. The test result acquisition unit in the test terminal further acquires an alternative answer from a caregiver for a plurality of questions for evaluating the daily life function of the subject,
The inspection result transmission unit in the inspection terminal transmits including the answer acquired by the inspection result acquisition unit,
The inspection result receiving unit in the inspection server receives including the answer,
The inspection result storage unit in the inspection server stores the answer including the answer,
The inspection server is
A correlation analysis unit for obtaining a correlation between the test score and the daily life function using the test score stored in the test result storage unit and the answer;
A correlation storage unit for storing the correlation between the test score obtained by the correlation analysis unit and the daily life function, and a comment for the subject whose daily life function corresponding to the question has decreased,
Further comprising
The prediction data transmission unit in the inspection server transmits the correlation and the comment stored in the correlation storage unit,
The prediction data receiving unit in the inspection terminal receives the correlation and the comment,
The prediction result generation unit in the inspection terminal refers to the correlation and the comment, and extracts a comment corresponding to a daily life function that correlates with a future inspection score of the target subject,
The cognitive function according to any one of claims 1 to 3, wherein the prediction result display unit in the inspection terminal displays a comment extracted by the prediction result generation unit as a comment on a daily life function decline predicted in the future. Prediction system. The inspection server is
Cognition of calculating the relationship between the elapsed time from the onset and the test score by nonlinear regression analysis of the test score for a large number of subjects stored in the test result storage unit and the elapsed time from the onset to the test date Function transition analysis unit,
An analysis result transmission unit that transmits the relationship between the elapsed time from the onset calculated by the cognitive function transition analysis unit and the test score to the test terminal;
Further comprising
The inspection terminal is
An analysis result receiving unit for receiving the relationship between the elapsed time from the onset and the test score transmitted from the test server;
An analysis result display unit for displaying a relationship between an elapsed time from the onset received by the analysis result receiving unit and a test score;
The cognitive function prediction system according to any one of claims 1 to 4, further comprising:   The cognitive function prediction system according to claim 1, wherein the cognitive function test is ADAS-Jcog. The inspection server is
A test content storage unit that stores test information including a method for evaluating the response of the subject in the cognitive function test, and score data for converting the response of the subject in the cognitive function test into a score,
An inspection content transmission unit for transmitting the inspection information and score data stored in the inspection content storage unit to the inspection terminal;
Further comprising
The inspection terminal is
Inspection content receiving unit for receiving the inspection information and score data transmitted from the inspection server,
An examination content display unit for displaying examination information received by the examination content receiving unit;
Further comprising
The test result acquisition unit in the test terminal scores the response result of the subject in the cognitive function test performed according to the test information displayed on the test content display unit, based on the score data received by the test content receiving unit And the cognitive function prediction system of any one of Claims 1-6 which acquires the obtained scoring result as a test | inspection score. The test information stored in the test content storage unit of the test server is prior knowledge for performing the test for each test item included in the cognitive function test, precautions for performing the test, and medical engagement in performing the test. Including the behavior that the person should do and how to evaluate the response of the subject in the test,
The test result acquisition unit in the test terminal is configured to obtain a response result of the subject for each test item included in the cognitive function test evaluated by a medical worker according to the test information displayed on the test content display unit. The cognitive function prediction system according to claim 7, wherein the content receiving unit scores based on the score data received, and obtains the obtained scoring result as an examination score.

Images