JP6864460B2 - Disease prediction device - Google Patents

Description

The present invention relates to a disease prediction method, a disease prediction program, a disease prediction device, and a disease prediction program for estimating diseases, disorders, etc. that the user will develop in the future from the activity information of the user.

Patients with dementia such as Alzheimer's may feel drowsy during the day or have reduced activity due to insomnia or night awakening. The reason for this is that as a symptom of dementia, a decrease in the amount of melatonin, which is a substance in the brain, occurs. Melatonin is a substance that regulates the rhythm of the body, and its secretion is controlled when a person is exposed to the morning sun. In patients with dementia, the time when the concentration of secreted melatonin peaks varies widely, and the timing of secretion becomes different from that in healthy subjects.
6 (a) and 6 (b) are both diagrams showing the concentration of melatonin in serum secreted on the vertical axis and the time of secretion on the horizontal axis, and FIG. 6 (a) is a healthy subject. belongs to. FIG. 6 (b) shows patients who developed Alzheimer's disease (Source: Mishima K, Tozawa T, Satoh K et al. Psychiatry 1999; 45: 417-421.). As is clear from FIGS. 6 (a) and 6 (b), in healthy subjects, the melatonin concentration increases from the evening, peaks at midnight, and decreases toward early morning. These changes in melatonin allow healthy individuals to act according to the rhythm of sleeping at night and waking up the next morning.

In contrast, melatonin levels in Alzheimer's patients change less during the day than in healthy individuals. For this reason, Alzheimer's patients are liable to have problems such as day-night reversal in their behavior because the time zone of the day does not match the timing of sleep and awakening.

By the way, known techniques for predicting a disease from the behavior of a subject are described in, for example, Patent Document 1 and Patent Document 2. In the method for evaluating the risk of senile disorder described in Patent Document 1, a sheet-type pressure sensor is used to detect the walking stride and angle of the subject, and the risk of senile disorder of the subject is predicted based on the detected value. Further, the bioanalytical apparatus described in Patent Document 2 attaches an acceleration sensor or the like to the subject and measures the activity in time series. Then, the measurement data is divided and analyzed, and the state of the subject is evaluated based on the relationship between the analysis result and the behavior and psychological state of the subject.

Japanese Unexamined Patent Publication No. 2013-255786 Japanese Unexamined Patent Publication No. 2008-67892

However, the walking experiment performed by attaching the sensor to the subject needs to be performed in a facility or the like equipped with equipment, which is a burden on the subject. For this reason, such an experiment is not suitable for performing it on a daily basis to detect a disorder in the rhythm of the subject's life at an early stage. In addition, the relationship between the subject's behavior and activity information such as heart rate is to detect signs of sudden motion sickness, panic disorder, and illness, and to detect disturbances in daily life rhythms. It is considered unsuitable.
The present invention has been made in view of the above points, and is a disease prediction method and a disease prediction program capable of predicting the onset of a disease from a disorder of daily life rhythm with a small load on the subject for measurement. , Disease Predictor and Disease Prediction Program.

In order to solve the above-mentioned problems, the disease prediction method of the present invention provides at least subject activity information regarding the status of activities performed by a plurality of subjects acquired in advance and the time zone in which the activities were performed, and the body of the subject. Diseases that the user develops by applying the user activity information regarding the status of the activity performed by the user and the time zone during which the activity was performed to the disease-related information indicating the relationship with the subject state information regarding the state. Alternatively, it is characterized by including a disease prediction step for predicting a disorder.

In addition, the disease prediction program of the present invention has at least subject activity information regarding the status of activities performed by a plurality of subjects acquired in advance and the time zone during which the activities were performed, and subject status information regarding the physical condition of the subject. By applying the user activity information regarding the status of the activity performed by the user and the time zone in which the activity was performed to the disease-related information indicating the relationship between the above and the above, the disease or disorder that the user develops is predicted. It is characterized by including a disease prediction step.

In addition, the disease prediction device of the present invention has at least subject activity information regarding the status of activities performed by a plurality of subjects acquired in advance and the time zone during which the activities were performed, and subject state information regarding the physical condition of the subject. By applying the user activity information regarding the status of the activity performed by the user and the time zone in which the activity was performed to the disease-related information indicating the relationship between the above and the above, the disease or disorder that the user develops is predicted. It is characterized by having a disease prediction unit.

In addition, the disease prediction system of the present invention includes an activity information acquisition device that acquires user activity information regarding the status of activities performed by the user and the time zone in which the activity was performed, and the user from the activity information acquisition device. It is characterized by including an input unit for inputting activity information and a disease prediction unit for predicting a disease or disorder that the user develops based on the user activity information.

The present invention described above provides a disease prediction method, a disease prediction program, a disease prediction device, and a disease prediction program that can estimate the onset of a disease from a disorder of daily life rhythm while the load on the subject for measurement is small. Can be provided.

It is a figure for demonstrating the disease prediction apparatus of one Embodiment of this invention. 2 (a) and 2 (b) are diagrams illustrating the configuration of the disease prediction system according to the embodiment of the present invention. It is a figure which shows the example which showed the INDEX of a disease for a plurality of diseases by a pentagonal graph in one Embodiment of this invention. It is a figure exemplifying the information which teaches the user how to judge dementia INDEX in one Embodiment of this invention. It is a flowchart for demonstrating the disease prediction method of one Embodiment of this invention. FIG. 6A is a graph showing the concentration of melatonin in the serum of healthy subjects. FIG. 6B is a graph showing the serum melatonin concentration of a patient who developed Alzheimer's disease.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings. In all drawings, similar components are designated by the same reference numerals, and duplicate description will be omitted as appropriate.

[Disease prediction device and disease prediction system]
(Disease prediction device)
FIG. 1 is a diagram for explaining the disease prediction device 10 of the present embodiment. The disease prediction device 10 is a disease that the user develops based on the status of the activity performed by the user of the disease prediction device (hereinafter, simply referred to as “user”) and the activity information regarding the time zone in which this activity was performed. Alternatively, it is provided with a disease prediction unit 132 that predicts a disorder (hereinafter, simply referred to as “disease, etc.”). Forecast of illness, etc., at least, the status of the activity performed by a plurality of subjects (hereinafter, simply referred to as "subjects") acquired in advance, the subject activity information regarding the time zone in which this activity was performed, and the subject's body. It is performed by applying the user activity information regarding the status of the activity performed by the user and the time zone during which the activity was performed to the disease-related information indicating the relationship with the subject state information regarding the state.
Here, the "activity" refers to the movement of the body performed regardless of the user's consciousness or unconsciousness. Activities that the user is conscious of include, for example, walking and swimming. Activities performed unconsciously include, for example, turning over, pulsating the heart, and maintaining posture in a sitting or standing position (resting state). Disease refers to physical abnormalities such as heart failure and cancer. In this embodiment, an endogenous substance that is particularly susceptible to lifestyle-related effects such as depression and diabetes is targeted. Disability, if not illness, refers to physical problems caused by aging and the like. In this embodiment, those caused by physical deterioration such as low back pain, knee pain and falls are targeted. A subject is a person who provides activity information and status information for creating disease-related information, and a user is a person who uses a disease prediction method using this disease-related information.
In addition, the "activity status" in the present embodiment means the amount of activity or activity intensity of the activity performed by the subject or the user during a specific time zone of the day. The "activity amount" is represented by the number and time of movements, the frequency of appearance of a certain intensity of activity, and the like. "Intensity of activity" is expressed by the force applied to the body during movement. "Subject status information" is information such as a disease obtained by actually inspecting or diagnosing a subject. The prediction of the disease or the like that the user develops is evaluated by INDEX such as dementia INDEX (index) and depression INDEX obtained by applying the user's activity information to the disease-related information. The INDEX of each disease is a numerical value indicating the degree of signs that the user develops the disease.
The disease-related information will be described in detail later.

The disease prediction device 10 includes an acceleration sensor 11 that acquires user activity information in order to predict a disease or the like. Further, it includes a control unit 13, a display unit 15, a timer 17, a memory 19 for storing related information, and an input unit 21 for performing processing based on user activity information. The input unit 21 may be any of a keyboard, a pen tab, a touch panel, and a voice input device.
(Acceleration sensor)
The acceleration sensor 11 is an acceleration sensor that detects acceleration in the three-axis directions of x, y, and z, and detects the acceleration applied to the disease prediction device 10 and the user with the y-axis direction as the gravity direction. As the acceleration sensor 11, for example, Lifecoder EX (registered trademark) manufactured by Suzuken Co., Ltd., Lifecoder PLUS, and the like are used. In the present embodiment in which activity information is acquired using the acceleration sensor 11, the walking motion of the user is measured as an activity. When measuring walking, the acceleration sensor 11 is attached near the waist of the user.
In this embodiment, activities other than walking may be adopted for acquiring activity information. As activities other than walking, for example, movement of the user's wrist can be considered. When measuring the number of times the wrist has moved or the speed, it is conceivable to wear the acceleration sensor 11 on the user's wrist. If the acceleration sensor 11 is attached to the wrist, it is possible to measure activities such as eating and drinking and operating the device while the user is sitting.

(Control unit)
The control unit 13 includes an activity information creation unit 131, a disease prediction unit 132, and a message creation unit 133. The activity information creation unit 131 inputs acceleration data related to the acceleration acquired by the acceleration sensor 11. Then, the number of walking steps of the user and the intensity thereof (hereinafter, also referred to as "activity intensity") are detected from the acceleration data. The activity intensity may be detected by, for example, the "speed" of walking.
Further, the activity information creation unit 131 creates activity information by associating the number of steps or the activity intensity with the time information acquired from the timer 17. Such a control unit 13 is a so-called computer, and has a known configuration such as a CPU (Central Processing Unit) (not shown), a memory for storing data, and a working memory. The activity information creation unit 131, the disease prediction unit 132, and the message creation unit 133 each include a hardover and a program that operate on a computer and exert a predetermined function.
The activity information creation unit 131 reads the acceleration signal output from the acceleration sensor 11 and calculates the amount of activity and the activity intensity by various arithmetic processes as described below. In this embodiment, the number of walking steps is measured as the amount of activity by the number of times acceleration is applied. In addition, the intensity of activity is defined as the walking speed, and the intensity of acceleration added to the speed is measured.

The activity information creation unit 131 may extract an acceleration component in the gravity direction from the acceleration signal acquired by the acceleration sensor 11. Various types of arithmetic processing for calculating the number of steps and speed from the acceleration signal have been proposed so far, and are not particularly limited. An example is shown below.
First, the activity information creation unit 131 converts the acceleration signal sequentially output from the acceleration sensor 11 into a voltage signal, digitally converts it, removes noise, synthesizes triaxial acceleration, extracts gravity direction components, and the like. Generate acceleration data. In this embodiment, the acceleration means that the influence of the gravitational acceleration of the earth is removed. That is, the walking user is loaded with an acceleration (for example, 1.2 G) exceeding 1 G (G is the gravitational acceleration) together with the gravitational acceleration. The activity information creation unit 131 acquires a value (0.2G in this case) obtained by excluding the influence of the gravitational acceleration from the measured acceleration as the acceleration.

When measuring the number of steps, the activity information creation unit 131 compares the acceleration data with the threshold value and measures the walking of the user when the acceleration value exceeds the threshold value. The number of times the walking is measured is counted as the number of steps. When measuring the activity intensity of walking, the activity information creation unit 131 compares the acceleration data with the threshold value to determine the magnitude of the acceleration, and calculates the activity intensity based on the determination result. The threshold value is stored in advance in the activity information creation unit 131. The activity information creation unit 131 constantly measures the acceleration from the user who is active with the acceleration sensor 11 attached, and the activity intensity due to the number of steps or walking is momentarily measured at predetermined time intervals (hereinafter referred to as intensity determination intervals). Is calculated.
In the present embodiment, the activity information creation unit 131 classifies the measured acceleration data into 11 stages of, for example, 0 to 9 and 0.5 based on the magnitude of the measured acceleration data. The classified acceleration data is stored in the above-mentioned data storage memory with the data measured during the intensity determination interval as one unit.

The intensity determination interval can be 1 second or more and 10 seconds or less, for example, 4 seconds or 6 seconds. By setting the intensity determination interval to 10 seconds or less, the activity status of the user can be analyzed in detail.
The activity information creation unit 131 generates one or a plurality of acceleration data based on the acceleration signal from the acceleration sensor 11 at each intensity determination interval, and determines the magnitude of the acceleration. Preferably, it is preferable to generate a plurality of acceleration data for each intensity determination interval and determine the magnitude of each acceleration. In this case, the sampling interval at which the acceleration sensor 11 acquires the acceleration signal is set shorter than the above-mentioned intensity determination interval, and more preferably, the acceleration sensor 11 measures the acceleration signal a plurality of times within the time of the intensity determination interval. Set to. Therefore, by setting the intensity determination interval to 1 second or more, many acceleration signals can be measured by the acceleration sensor 11 within the time of the intensity determination interval without excessively shortening the sampling interval of the acceleration sensor 11. ..

The activity information creation unit 131 determines the activity intensity of the user within the time of the intensity determination interval based on the number of occurrences of each determined acceleration magnitude. As a result, it is possible to prevent the activity intensity from being excessively determined due to the momentary detection of a large acceleration. Further, even if the magnitudes of the detected accelerations are the same, the activity intensity can be distinguished into different intensity stages based on the magnitude of the occurrence frequency of the accelerations.
The activity information creation unit 131 may determine the activity intensity by averaging the accelerations counted within the time of the intensity determination interval. This prevents excessively high activity intensity from being measured due to the large acceleration recorded at the moment when the user stands up or falls down. In this way, the activity information creation unit 131 can generate one activity intensity for each intensity determination interval based on a plurality of acceleration data acquired within the time of the intensity determination interval.

The disease prediction unit 132 reads the disease-related information from the memory 19 for storing the related information. Then, the activity information created by the activity information creation unit 131 is applied to the disease-related information to predict the disease or the like that the user develops.
The message creation unit 133 compares the activity information of the user with the activity information of the subject and the height, weight, and age, and includes advice information suggesting the status of the activity to be performed by the user and the time zone in which this activity should be performed. Create a message. The message including the advice information created by the message creation unit 133 is output to and displayed on the display unit 15. The display unit 15 may be a known display device. Further, the present embodiment is not limited to the configuration in which the message is displayed on the display unit 15 as text, and the user may be notified by voice or light of the time zone and time interval in which the activity should be performed. ..

Such a disease prediction device 10 of the present embodiment may be configured by using, for example, a smartphone having a pedometer function. When the disease prediction device 10 is configured by using a smartphone, a known configuration provided in the acceleration sensor 11, the timer 17, the related information storage memory 19, the input unit 21, and the display unit 15 can also be used.

(Disease prediction system)
Further, the present embodiment is not limited to the configuration of the disease prediction device 10 described above, and the activity information acquisition device for acquiring the activity information and the disease prediction for predicting the disease by arithmetically processing the acquired activity information. It can also be separated from the device.
2 (a) and 2 (b) are diagrams exemplifying the configuration in which the activity information acquisition device and the disease prediction device are separated. In the present embodiment, the configuration in which the activity information acquisition device and the disease prediction device are separated is also referred to as a “disease prediction system”. The configuration shown in FIG. 2A shows an example in which a disease prediction system is constructed by an activity information acquisition device 30 and a disease prediction device 50 that predicts a disease using the activity information. The activity information acquisition device 30 includes an output interface (I / F) 31 in addition to the acceleration sensor 11, the activity information creation unit 131, and the timer 17. The disease prediction device 50 includes an input interface (I / F) 51 in addition to the disease prediction unit 132, the message creation unit 133, the related information storage memory 19, the input unit 21, and the display unit 15.

The output I / F 31 and the input I / F 51 may input activity information from the activity information acquisition device 30 to the disease prediction device 50 using a data recording medium. Further, the activity information may be exchanged by wire or wirelessly, or the activity information may be transmitted / received by infrared communication.
Further, as shown in FIG. 2B, the present embodiment includes an activity information acquisition device 20 including an acceleration sensor 11, an activity information creation unit 131, and a timer 17, a disease prediction unit 132, and a message creation unit 133. A computer system may be constructed by connecting the disease prediction device 40 provided with the related information storage memory 19, the input unit 21, and the display unit 15 to the network N.

When the present embodiment is divided into the activity information acquisition device 30 and the disease prediction device 50 as shown in FIG. 2A, the activity information acquisition device worn by the user on the body is miniaturized and the activity information of the user is used. The load on the acquisition device can be reduced. Further, since the disease prediction unit 132 and the message creation unit 133 can be used as highly accurate devices, it is possible to predict the disease or the like that the user develops with high accuracy. In addition, the type of predicted disease and the content of the message can be diversified. Further, as shown in FIG. 2B, if the acquired activity information is sequentially transmitted to the disease prediction device, a message can be transmitted to the user in real time.

[Disease prediction method]
Next, the disease prediction method of the present embodiment will be described.
The disease prediction method of the present embodiment relates to at least the disease-related information indicating the relationship between the activity information of the subject and the state information of the subject, the status of the activity performed by the user, and the time zone during which the activity was performed. It includes a process of predicting the disease, etc. that the user will suffer by applying the activity information of the user.
Further, in the disease prediction method of the present embodiment, in the above step, the disease of the user is further predicted based on the information regarding at least one of the height, weight and age of the user.
In the present embodiment, prior to the explanation of such a disease prediction method, first, disease-related information will be described.

(Disease-related information)
In this embodiment, disease-related information will be described by taking as an example the relationship between the number of steps, intensity, and walking time zone of walking and the state information of dementia and depression. The inventor of the present invention obtained activity information of each person from 50 or more subjects, preferably 100 or more subjects. For the acquisition of activity information, the walking of the subject was used as an index of activity, the number of steps of walking was used as "activity amount", and the speed of walking was used as "activity intensity". Further, the activity information was acquired by wearing the activity information acquisition device 30 shown in FIG. 2A on the waist while the subject was awake for 28 days. Further, the activity information was acquired by walking every 3 hours as one unit. Therefore, in the present embodiment, the number of steps and the intensity of walking of the subject for 28 days are recorded every 3 hours in association with the time zone.
The disease estimation method of the present embodiment is not particularly limited in the age of the applicable subject. This is because depression also develops in young people such as those in their twenties and thirties. However, when predicting dementia as a disease, the age of the subject is preferably 60 years or older, more preferably 65 years or older, and even more preferably 70 years or older. Further, as the subject, for example, a person who tends to have irregular activity hours such as a medical worker, a firefighter, and a police officer is not preferable.

Table 1 is a table showing disease-related information about depression and dementia in the present embodiment. As shown in Table 1, all the disease-related information of the present embodiment is expressed by a linear equation using parameters including activity information related to walking.
Hereinafter, a method for creating disease-related information according to this embodiment will be described.
In this embodiment, disease-related information is created by performing the following steps (A) to (E).
(A) Obtain one or more data on the subject's physique such as age, gender, height and weight (BMI).
Such a step is executed, for example, by inputting by an operator from the input unit 21 of the disease prediction device 50 shown in FIG. 2A.

(B) Acquire activity information such as the number of steps and the intensity of walking every 3 hours regarding the subject's walking.
The activity information is created by the activity information creation unit 131 processing the acceleration data acquired by the acceleration sensor 11 of the activity information acquisition device 30 shown in FIG. In this embodiment, the parameters such as height applied to the disease-related information together with the activity information created by the activity information creation unit 131 are defined as 44 types as follows. When the parameter is an average value, this average value is an average value over the entire measurement period (28 days) unless otherwise specified.

When the exercise is walking (including running), the present embodiment defines 44 types of parameters shown in the following (1) to (10). Then, using this parameter, disease-related information for predicting the disease that the subject develops is created.

(1) 1-8; Average value of the number of steps in each time zone The number of steps (Ns _3-6 ) in each set time zone (for example, 3-6 o'clock) for 3 hours is measured. _{Then, the average value (Ns 3-6ave} ) of the number of steps in each time zone for 28 days is calculated and used as the average value of the number of steps in each time zone.
Ns _3-6ave , Ns _6-9ave , Ns _9-12ave , Ns _12-15ave , Ns _15-18ave , Ns _18-21ave , Ns _21-24ave , Ns _24-3ave

(2) 9 to 16; average value of activity intensity in each time zone The activity intensity (Np _3-6 ) is measured for each set time zone (for example, 3-6 o'clock) for 3 hours. Since the activity intensity changes during 3 hours, the activity intensity within the time determined by the activity information creation unit 131 that the user is walking is calculated for each intensity determination interval, and the value obtained by averaging them over time is calculated. The average activity intensity during that time period. Then, the average value (Np _3-6ave ) for 28 days is calculated for the average activity intensity of each time zone, and the average value of the activity intensity of each time zone is used.
Np _3-6ave , Np _6-9ave , Np _9-12ave , Np _12-15ave , Np _15-18ave , Np _18-21ave , Np _21-24ave , Np _24-3ave

(3) 17-27; average value of the frequency of walking with intensity in each stage (0 to 9 and 0.5) The number of walking performed in one day is calculated for each intensity (Nf ₀ , Nf _0.5). , Nf ₁ , Nf ₂ , Nf ₃ , Nf ₄ , Nf ₅ , Nf ₆ , Nf ₇ , Nf ₈ , Nf ₉ ) are counted daily. _{Then, the average value (Nf 0ave} ) for 28 days is calculated for the number of walks of each intensity, and the average value of the frequency of walking of each intensity is used.
Nf _0ave , Nf _0.5ave , Nf _1ave , Nf _2ave , Nf _3ave , Nf _4ave , Nf _5ave , Nf _6ave , Nf _7ave , Nf _8ave , Nf _9ave

(4) 28; Coefficient of variation of the number of steps The standard deviation Ns _sd of the number of steps for 28 days is calculated. _{In addition, the average value (Ns ave} ) obtained by averaging the measured number of steps over 28 days is calculated. Then, the following value obtained by dividing the standard deviation Ns _sd by the average value (Ns _ave ) is used as the coefficient of variation of the number of steps.
Ns _sd / Ns _ave x 100

(5) 29; Coefficient of variation of wearing time Calculate _{the standard deviation Nisd of wearing time for 28 days. In addition, the average value (Niave} ) obtained by averaging the measured wearing time over 28 days is calculated. Then, the following value obtained by dividing the standard deviation Ni _sd by the average value Ni _ave is used as the mounting time coefficient of variation.
Ni _sd / Ni _ave x 100

(6) 30; calculating a standard deviation _{Np sd} of activity intensity of activity intensity variation coefficient 28 days. In addition, the average value (Np- _ave ) obtained by averaging the measured activity intensity over 28 days is calculated. Then, the following value obtained by dividing the standard deviation Np _sd by the average value Np _ave is used as the coefficient of variation of activity intensity.
Np _sd / Np _ave × 100

(7) 31-41; Coefficient of variation of activity intensity frequency in each stage (0-9 and 0.5) Frequency of walking at each stage (0-9 and 0.5) intensity (Nf ₀ , Nf _0.5 , Nf ₁ , Nf ₂ , Nf ₃ , Nf ₄ , Nf ₅ , Nf ₆ , Nf ₇ , Nf ₈ , Nf ₉ ) are counted daily. In addition, the standard deviation of the counted frequency is calculated for each intensity (Nsd ₀ , Nsd _0.5 , Nsd ₁ , Nsd ₂ , Nsd ₃ , Nsd ₄ , Nsd ₅ , Nsd ₆ , Nsd ₇ , Nsd ₈ , Nsd _9). ). Then, the standard deviations (Nsd ₀ , Nsd _0.5 , Nsd ₁ , Nsd ₂ , Nsd ₃ , Nsd ₄ , Nsd ₅ , Nsd ₆ , Nsd ₇ , Nsd ₈ , Nsd ₉ ) are shown in (3). 28-day average value (Nf _0ave , Nf _0.5ave , Nf _1ave , Nf _2ave , Nf _3ave , Nf _4ave , Nf _5ave , Nf _6ave , Nf _7ave , Nf _6ave , Nf _{7ave, Nf 8ave, Nf 8ave, Nf 8ave} It is the coefficient of variation of the frequency of activity intensity at each stage.
Nsd ₀ / Nf _0ave x 100, Nsd _0.5 / Nf _0.5ave x 100, Nsd ₁ / Nf _1ave x 100, Nsd ₂ / Nf _2ave x 100, Nsd ₃ / Nf _3ave x 100, Nsd ₄ / Nf _4a 100, Nsd ₅ / Nf _5ave x 100, Nsd ₆ / Nf _6ave x 100, Nsd ₇ / Nf _7ave x 100, Nsd ₈ / Nf _8ave x 100, Nsd ₉ / Nf _9ave x 100
However, the above Nsd ₀ and Nf _0ave exclude the information of the time zone when the subject is not wearing the activity information acquisition device 30.
(8) 42; Age (9) 43; Height (10) 44; Weight

Further, in this embodiment, not only the measured activity information itself but also the activity amount of the activity performed by the subject or the degree of dispersion of the activity intensity is used as the parameter used for the prediction of the disease. The degree of dispersion of the activity amount or the activity intensity is, for example, a parameter indicating "variation" within the measurement period of the activity amount or the activity intensity, and is represented by the above-mentioned coefficient of variation in the present embodiment.
Further, in the present embodiment, information on the distribution of the amount of activity during the day of the activity can also be used for estimation as the activity information. The distribution of the amount of activity is represented by, for example, "morning activity time ratio", "daytime activity time ratio", "night activity time ratio", and the like. The morning activity time ratio indicates the ratio of the number of walking steps taken from 3:00 to 9:00 to the number of steps per day. The daytime activity time ratio indicates the ratio of the number of walking steps taken from 9:00 to 18:00 to the number of steps per day. The night activity time ratio indicates the ratio of the number of walking steps taken from 18:00 to 3:00 to the number of steps per day.

Further, among the above parameters, the average value of the activity intensity of (2) can be used for estimating the walking speed of the user. The activity intensity is estimated by applying the average value of the activity intensity to the multiple regression equation. The multiple regression equation for estimating the walking speed is as follows.
(Number 1)
(male)
Estimated walking speed (Km / h) = 0.862 x (average activity intensity) + 0.035 x (height (cm))-4.705 ... Equation (1)
(Number 2)
(Woman)
Estimated walking speed (Km / h) = 0.684 x (average activity intensity) +1.851 ... Equation (2)
(Source: Naoto Takayanagi et al., "Study on Daily Walking Speed and ADL Decrease Using Activity Meter" Health Index. 2014; 61 (4): 15-20)

The numerical value such as the activity intensity "5" of the present embodiment is, for example, the maximum value of the activity intensity at which the user is considered to ingest the oxygen of the maximum oxygen uptake defined by METs, and the minimum value of the oxygen intake at rest. It can be determined by dividing the space between them.
In terms of METs, for example, 3METs walks on flat ground at 67 m / min, 3.3METs walks on flat ground at 81 m / min, 3.8METs walks on flat ground at 94 m / min, and 4METs walks on flat ground at 95 m to 100 m / min. It stipulates that walking on flat ground in minutes and 5METs at 107 m / min on flat ground (Source: Ministry of Health, Labor and Welfare "Exercise Guidelines for Health Promotion 2006").
The regression analysis of this embodiment was performed using commercially available software for statistical calculation. Multivariate regression analysis and logistic regression analysis are used for regression analysis. In the regression analysis, the parameters used for the explanatory variables are selected by performing the variable increase method, the variable decrease method, or both, and changing the combination to increase the correct answer rate. Further, the coefficients attached to the parameters are appropriately selected.

However, the parameters including the activity information of the present embodiment are not limited to the examples described above. For example, the activity information is not limited to the one measured as one unit every three hours, and may be measured in any time as a unit. The average value of the parameters may be the average value per unit time of measurement, the average value per day, or the average value of the measurement period over a plurality of days.

(C) Obtain information regarding the presence or absence of symptoms of depression or dementia of the subject.
Such a process is performed by interviewing or inspecting the subject. That is, it can be said that this step actually measures the presence or absence of depression and dementia in the subject.
(D) Logistic regression analysis was performed using the presence or absence of depression or dementia actually measured in (C) as the objective variable, and at least one of the activity information acquired in (B) or parameters such as height and weight were used as explanatory variables. Calculate the including regression equation.
In the step (D), both the data of the person who clearly has the symptoms of dementia and depression and the data of the person who does not have the symptoms of dementia are extracted from the result of the actual measurement performed in (C). Specifically, as a result of step (C), four groups of "none", "current illness (without medication)", "current illness (with medication)" and "previous" were extracted from the subjects, and "none". The subjects who have been identified as "" are clearly asymptomatic. In addition, subjects with "current illness (without medication)" and "current illness (with medication)" are considered to have obvious symptoms.

(E) In the steps (D), the value of the objective variable obtained from the activity information of the "symptomatic" subject is "1", and the value of the objective variable obtained from the activity information of the "symptomatic" subject is "0". Create symptom presence / absence data as. Then, regression analysis is performed using one or more activity information and height of the subject related to the symptom presence / absence data as explanatory variables. The regression analysis was performed using commercially available software for statistical calculation. Multivariate regression analysis and logistic regression analysis are used for regression analysis. In the regression analysis, the parameters used for the explanatory variables are selected by performing the variable increase method, the variable decrease method, or both, and changing the combination to increase the correct answer rate. Further, the coefficients attached to the parameters are appropriately selected.

Table 1 is a table specifically showing disease-related information on depression and dementia obtained by using the above parameters. In Table 1, the numbers in parentheses () after the “;” in the relevant information of the subject correspond to the numbers in parentheses () of the above parameters.
Depression, disease-related information about dementia, as shown in Table 1, in both the value of the model chi ² test with the measured value of 0.01 or less, it can be seen that significant. In addition, a high hit rate of 82.8% for depression and 97.2% for dementia was obtained. The disease-related information shown in Table 1 is common to both men and women.

Tables 2, 3, 4 and 5 show disease-related information, Tables 2 and 3 show related information about men, and Tables 4 and 5 show related information about women. ing. In any of Tables 2 to 5, "backache", "knee pain", and "urinary incontinence" are listed as disorders. In addition, as diseases, "depression", "dementia", "diabetes", "hypertension", "hyperlipidemia", "osteoporosis", "osteoarthritis" and "spine disease" are mentioned.
In this embodiment, the user's state information can be obtained by inputting the user's parameters in the parentheses "()" of each disease-related information in Tables 1 to 5 and solving the calculation formula. Further, the disease-related information of the present embodiment is, of course, not limited to the form of the linear expression as shown in Tables 1 to 5. The disease-related information may be any information as long as it predicts the disease by inputting the user's activity information, height, etc., and may be, for example, an arithmetic expression other than the linear expression. It may be a table.

(message)
Next, an example of a message including advice information created based on the prediction result of the user's disease will be described. In the present embodiment, a message including advice on activities to be performed by the user is created based on the INDEX of the disease obtained as a result of applying the activity information to the above-mentioned disease-related information. The advice is information that suggests the amount, intensity, time zone, etc. of the activity to the user, and the message is the advice information with necessary information added as appropriate. Examples of the information added to the advice information include those shown in FIGS. 3 and 4.
FIG. 3 is a diagram showing an example in which the INDEX of diseases for a plurality of diseases is shown by a pentagonal graph. In the example shown in FIG. 3, five vertices of dementia INDEX, depression INDEX, osteoarthropathy INDEX, osteoporosis INDEX, and knee pain INDEX are shown at each of the five vertices of the pentagon. The value of each INDEX increases as the distance from the center of the pentagon increases. In the example shown in FIG. 3, the prediction results for a plurality of diseases and the like are shown, and the balance can be grasped by the user.

FIG. 4 is a diagram illustrating information for teaching the user how to determine dementia INDEX. In this embodiment, the user's dementia INDEX is used as advice information, and for example, the information shown in FIG. 4 is added to the advice information to create a message. In this way, the user can grasp the degree of possibility that he / she will develop dementia from his / her dementia INDEX.
For example, when the user's dementia INDEX is in the range of 0.5 to a predetermined threshold value (indicated by "○" in the figure), the user himself / herself has dementia from the "dementia risk" (left column). It is determined that the possibility of developing dementia is about 60%. "Dementia risk level" (center column) shows the possibility of developing dementia in a pentagonal graph. At the dementia risk level, the area indicated by "2" indicates that the possibility of developing dementia is 60%, and the area indicated by "1" indicates that the possibility of developing dementia is 0%, "3". The areas shown in are 80% of the probability of developing dementia. At the dementia risk level, for example, the area corresponding to the user's dementia prediction may be colored and displayed in a predetermined color. Further, the example of FIG. 5 further includes a “dementia risk rank” (right column) in which the prediction of the onset of dementia of the user is evaluated at the levels of A, B, and C.

For example, when the user desires improvement of dementia, the message creation unit 133 sets the average value of the intensity of the activities performed from 3:00 to 6:00 so that the dementia INDEX shown in Table 1 becomes less than 0.5. Set the average value of the number of steps of walking performed from 18:00 to 21:00. Then, compare each set average value with the history of the user's activity information, and when the history of the intensity of the user's early morning activity is less than the set value, "Walk a little early in the morning. Create a message that includes advice such as "Let's go".

Further, in the present embodiment, the type of the disease or the like that the user desires to improve can be input in advance by the input unit 21 shown in FIGS. 1 and 2. The message creation unit 133 inputs the height, weight, and age of the user into the disease-related information shown in Tables 1 to 3, and calculates the number of steps and the intensity that the user should walk in response to the disease or the like. Information on body weight, height, and age may be registered in advance in the disease prediction device 50 using the input unit 21 when the user starts using the disease prediction method of the present embodiment, or may be arbitrarily registered from the input unit 21. It may be input at the timing of.
Further, in the present embodiment, the user can input, for example, the number of steps and the strength of walking performed in 3 hours from the input unit 21. In this way, when the number of steps or activity intensity to be increased exceeds the upper limit set by the user, the message creation unit 133 advises the user to walk with the number of steps or intensity set in the upper limit. Compose a message. According to such processing, even if the difference between the disease prediction information desired by the user and the current situation is relatively large, it is possible to reduce the possibility of developing the disease while suppressing the pain of the user by taking time. it can. Further, the message creation unit 133 creates a message such as "the sign of the onset of dementia has been reduced" when the possibility of developing the user's illness or the like is sufficiently low.

The above-mentioned message may be further printed on a paper medium and output, or may be transmitted as an e-mail or the like to a user's smartphone or the like. Such an e-mail may be sent to a third party (family, etc.) registered in advance in addition to the person himself / herself. Further, in the present embodiment, the activity information acquisition device 30 may also be provided with a display unit (not shown), and a message may be presented to both of them through the display unit of the activity information acquisition device 30. Further, in the present embodiment, when the activity information acquisition device 30 is connected to the disease prediction device 50, a message can be displayed on the display unit 15 of the disease prediction device 50.

(Disease prediction method)
FIG. 5 is a flowchart for explaining the disease prediction method of the present embodiment. Further, the flowchart of FIG. 5 shows an example performed by the activity information acquisition device 30 and the disease prediction device 50 shown in FIG. 2 (a). In this flowchart, it is assumed that the process of measuring the activity intensity for 3 hours as the activity information of the user and calculating the average value of the measured values is illustrated. Further, in the flowchart of FIG. 5, for the sake of simplicity of explanation, an example of acquiring one activity information once is shown. However, as shown in Tables 1 to 3, the present embodiment often acquires a plurality of types of activity information a plurality of times.
In the disease prediction method of the present embodiment, when the processing is started, the acceleration sensor 11 of the activity information acquisition device 30 starts acquisition of the user's activity information (step S501). It should be noted that such processing may be automatically started at a time corresponding to the disease that the user desires to predict, or may be started at a predetermined time determined in advance. ..
The “time according to the disease that the user wants to predict” refers to, for example, 6:00 or 18:00 when the user wants to predict the depression shown in Table 1.

When the acquisition of the user's activity information is started, the timer 17 continuously clocks until 3 hours have elapsed (step S502). The acceleration sensor 11 detects the acceleration applied to the user while the timer 17 is timing, and continues to output the activity intensity to the activity information creation unit 131. The activity information creation unit 131 calculates the average value of the activity intensity measured by the acceleration sensor 11 for 3 hours (step S503). In addition, the activity information creation unit 131 creates activity information by associating the calculated average value with the time zone in which the measurement was performed. The created activity information is sent to the disease prediction unit 132 via the output I / F31 and the input I / F51. The disease prediction unit 132 reads the disease-related information from the related information storage memory 19 and predicts the disease or the like that the user suffers from (step S504).

The message creation unit 133 determines, for example, the predicted user's illness or the like. Then, the number of walking steps and the intensity required to reduce the sign of the disease from the INDEX of the disease are set. Furthermore, advice information that advises the user on the number of steps and activity intensity is created from the difference between the set value and the walking history actually performed by the user. Then, for example, the information shown in FIGS. 3 and 4 is added to the advice information to create a message (step S505). The message is output to the display unit 15 (step S506).
The output of the message is not limited to the timing described in the flowchart shown in FIG. For example, as in the case of depression shown in Table 1, when activity information is acquired from activities performed twice, from 6:00 to 9:00 and from 18:00 to 21:00, the activities of the activities performed by 21:00. Predict disease from information. Then, a message created based on the predicted type of disease may be provided to the user at the start of the activity on the next day, for example.

At least steps S504 to S506 of the flowchart described above are processed by, for example, a program realized in the disease prediction device 50 configured as a computer.

The present embodiment described above predicts a user's illness based on the activities performed by the user. Therefore, this embodiment can predict the disease of the user without using a special device. In addition, since the present embodiment can automatically acquire activity information from the user, it is possible to make a habit of predicting a disease without imposing a burden on the user.
Further, in this embodiment, the disease is predicted by applying the activity information to the disease-related information in consideration of the measured value of the subject based on the principle of time ethology. Therefore, an estimated value having a high correlation with the measured value can be obtained.

Next, an embodiment of the present embodiment described above will be described.
The present inventors took 1,483 men and 1,598 women aged 70 to 93 as subjects, and obtained the state of depression and the state of dementia of each subject by interview. In addition, the present inventors acquired the height, weight, and age of the subjects, and calculated the BMI from the height and weight.
Next, the present inventors made each subject carry an acceleration sensor for 24 hours, and acquired the number of steps and intensity of walking of the subject as activity information. The activity information acquired by the accelerometer is output to a computer as a disease prediction device. The computer created disease-related information by performing logistic regression analysis using the acquired activity information and information such as the height of the subject. The disease-related information of this example is a linear regression equation for calculating px. By substituting px obtained by using this first-order regression equation into logpx / (1-px), the incidence rate of the subject x's disease or the like can be obtained, and this incidence rate takes a numerical value from 0 to 1.

The disease-related information about depression of this example prepared by the above processing is as follows. In this example, disease-related information is sought separately for men and women.
(Number 3)
Disease-related information (depression: male) = ^-3 x 10 -3 x (average number of walking steps taken from 21:00 to 24:00) -0.576 x (average value of activity intensity taken from 6:00 to 9:00) ) -0.65 ... Equation (3)
(Number 4)
Disease-related information (depression: female) = 10 ^-3 x (average number of walking steps performed from 3:00 to 6:00) -0.819 x (average value of activity intensity performed from 12:00 to 15:00) + 0. 162 x (age) +0.018 (ratio of daytime activity time)-12.585 ... Equation (4)

In addition, the disease-related information about dementia of this example calculated by the above processing is as follows. In this example, disease-related information on dementia was created only for women.
(Number 5)
Disease-related information (dementia: female) = -0.022 x (average number of walking steps performed from 21:00 to 24:00) -2.774 x (average value of activity intensity performed from 3:00 to 6:00) +0.29 × (age) -23.916 ... Equation (5)

When performing logistic regression analysis, the present inventors so that the correct answer rate (degree of agreement with the state information obtained by interview) of the calculation result (INDEX) of the disease-related information obtained from the regression equation is high. You have selected a combination of activity information and parameters for which related information is used. In this embodiment, the number of steps and intensity of walking of the user can be applied to the disease-related information to discriminate the signs of depression and dementia of the user, and the disease or the like can be predicted. In this example, the onset of depression or dementia is predicted when the depression INDEX or dementia INDEX is 0.5 or more and 1.0 or less, and the depression INDEX or dementia INDEX is 0.5 or less. It was predicted that the case would not develop depression or dementia. The correct answer rate for the above predictions in this example was 88.2% for male depression, 81.3% for female depression, and 96.5% for female dementia, all of which were high.

Furthermore, the present inventors, the reliability of the regression equation obtained in this example was examined by a model X ² test. In the model X ² test, significant probability p men and women both less than 0.05 for depression, significant probability even in women with dementia was less than 0.05. From this, all the regression equations obtained in this example have sufficient reliability.

The above-described embodiments and examples include the following technical ideas.
<1>
At least, disease-related information indicating the relationship between the status of activities performed by a plurality of subjects acquired in advance and the subject activity information regarding the time zone during which the activity was performed, and the subject status information regarding the physical condition of the subject. Includes a disease prediction step of predicting the disease or disorder that the user develops by applying the user activity information regarding the status of the activity performed by the user and the time zone in which the activity was performed. , Disease prediction method.
<2>
The disease prediction method according to <1>, wherein the disease prediction step further predicts the disease of the user based on information on at least one of the height, weight, and age of the user.
<3>
The disease prediction method according to <1> or <2>, wherein the subject activity information is information on the amount of activity or activity intensity of the activity performed by the subject during a specific time zone of the day.
<4>
The disease prediction method according to <1> or <2>, wherein the subject activity information is information on the daily dispersion of the activity amount or activity intensity of the activity performed by the subject.
<5>
Described in any one of <1> to <4>, further including a message creation step of creating a message including advice on activities to be performed by the user based on the result of prediction performed in the disease prediction step. Disease prediction method.
<6>
At least, disease-related information indicating the relationship between the status of activities performed by a plurality of subjects acquired in advance and the subject activity information regarding the time zone during which the activity was performed, and the subject status information regarding the physical condition of the subject. Includes a disease prediction step that predicts the disease or disorder that the user develops by applying the user activity information regarding the status of the activity performed by the user and the time zone in which the activity was performed. Disease prediction program.
<7>
At least, disease-related information indicating the relationship between the status of activities performed by a plurality of subjects acquired in advance and the subject activity information regarding the time zone during which the activity was performed, and the subject status information regarding the physical condition of the subject. It is characterized in that it is provided with a disease prediction unit that predicts the disease or disorder that the user develops by applying the user activity information regarding the status of the activity performed by the user and the time zone in which the activity was performed. Disease prediction device.
<8>
An activity information acquisition device that acquires user activity information regarding the status of activities performed by the user and the time zone in which the activity was performed, and
An input unit for inputting the user activity information from the activity information acquisition device, and
The disease prediction device according to <7>, comprising a disease prediction unit that predicts a disease or disorder that the user develops based on the user activity information.
A disease prediction system characterized by including.

10, 40, 50 ... Disease prediction device 11 ... Accelerometer 13 ... Control unit 15 ... Display unit 17 ... Timer 19 ... Related information storage memory 20, 30 ... Activity Information acquisition device 21 ... Input unit 31 ... Output I / F
51 ... Input I / F
131 ・ ・ ・ Activity information creation department 132 ・ ・ ・ Disease prediction department 133 ・ ・ ・ Message creation department

Claims (6)

At least, disease-related information indicating the relationship between the status of the activity performed by the plurality of subjects acquired in advance and the subject activity information regarding the time zone in which the activity was performed and the subject status information regarding the physical condition of the subject. It is equipped with a disease prediction unit that predicts the disease or disorder that the user develops by applying the user activity information regarding the status of the activity performed by the user and the time zone in which the activity was performed.
The activity is a physical movement of the subject or the user performed regardless of consciousness or unconsciousness.
The subject activity information includes information on the amount of activity and information on the intensity of the activity performed by the subject at a specific time of the day.
The user activity information, only contains the information about the information and the activity intensity on the activities the amount of activity that the user went to the particular time zone,
The specific time zone is a day divided into units of a predetermined length of time.
The disease prediction device, wherein the subject activity information and the user activity information include information on the activity intensity of the activity performed in the morning time zone and information on the activity amount of the activity performed in the night time zone. .. The disease prediction device according to claim 1, wherein the disease prediction unit further predicts the disease of the user based on information regarding at least one of the height, weight, and age of the user. The disease prediction device according to claim 1 or 2 , wherein the subject activity information further includes information on the daily dispersion of the activity amount or activity intensity of the activity performed by the subject. The disease according to any one of claims 1 to 3 , further comprising a message creation unit that creates a message including advice on activities to be performed by the user based on the result of prediction made by the disease prediction unit. Predictor. At least, disease-related information indicating the relationship between the status of the activity performed by the plurality of subjects acquired in advance and the subject activity information regarding the time zone in which the activity was performed and the subject status information regarding the physical condition of the subject. Includes a disease prediction step that predicts the disease or disorder that the user develops by applying user activity information about the status of the activity performed by the user and the time zone during which the activity was performed.
The activity is a physical movement of the subject or the user performed regardless of consciousness or unconsciousness.
The subject activity information includes information on the amount of activity and information on the intensity of the activity performed by the subject at a specific time of the day.
The user activity information, only contains the information about the information and the activity intensity on the activities the amount of activity that the user went to the particular time zone,
The specific time zone is a day divided into units of a predetermined length of time.
A disease prediction program in which the subject activity information and the user activity information include information on the activity intensity of the activity performed in the morning time zone and information on the activity amount of the activity performed in the night time zone. An activity information acquisition device that acquires user activity information regarding the status of activities performed by the user and the time zone in which the activity was performed, and
An input unit for inputting the user activity information from the activity information acquisition device, and
The disease prediction device according to claim 1, further comprising a disease prediction unit that predicts a disease or disorder that the user develops based on the user activity information.
A disease prediction system characterized by including.

Images