TWI862478B - Physical condition prediction method, computer and physical condition prediction program - Google Patents

Abstract

Provided is a physical condition prediction method, a physical condition prediction device, and a physical condition prediction program capable of predicting changes in the physical condition of a subject.

The body condition prediction method obtains the body movement data of the target person (step S6), continuously determines the sleep state of the target person based on the obtained body movement data (step S7), and predicts the change of the body condition of the target person based on the determined sleep state (step S9).

Description

Physical condition prediction method, computer and physical condition prediction program

Invention Field

This disclosure relates to a physical condition prediction method, a physical condition prediction device, and a physical condition prediction program for predicting the physical condition of a subject.

Invention background

In the past, there is a known method of using a physical condition management system to monitor the daily sleep state of the subject and perform appropriate physical condition management. The physical condition management system has: a sleep sensor installed on the subject; and an information terminal that analyzes or obtains and records the measurement data obtained by the sleep sensor (for example, refer to patent document 1).

The sleep sensor of Patent Document 1 analyzes the sleep state of the subject from the measurement data obtained by the sleep sensor and drives the display unit or speaker. In addition, in Patent Document 1, the sleep state of the subject determined by the sleep sensor is used to control electric curtains, audio equipment, lighting equipment, television, air conditioning equipment and bedding (electric bed or air mattress, etc.).

Prior art literature

Patent Literature

Patent document 1: Japanese Patent Publication No. 2013-150660

Summary of invention

However, the above-mentioned learning techniques do not take into account the need to predict changes in the physical condition of the target person, which requires further improvement.

This disclosure is an invention made to solve the above-mentioned problem, and its purpose is to provide a physical condition prediction method, a physical condition prediction device and a physical condition prediction program that can predict changes in the physical condition of a subject.

One aspect of the physical condition prediction method disclosed herein is to obtain physiological data of a subject, continuously determine the sleep state of the subject based on the obtained physiological data, and predict changes in the physical condition of the subject based on the determined sleep state.

According to the present disclosure, it is possible to predict changes in the physical condition of the target person based on the continuously determined sleep state. In addition, since changes in the physical condition of the target person are predicted, for example, when the target person is an elderly person or a patient with dementia, the care plan for the target person can be re-evaluated, and the target person can be cared for more efficiently.

1: Server

2: Motion sensor

3: Temperature sensor

4: Terminal device

5: Internet

11: Communications Department

12: Control Department

13: Memory Department

111: Body movement data acquisition unit

112: Physical status data acquisition unit

113: Infrared image acquisition unit

114: Forecast result sending department

121: Sleep determination unit

122: Physical condition analysis department

123: Body temperature determination unit

124: Physical condition prediction department

131: Sleep state storage unit

132: Physical status data storage unit

133: Physical condition prediction information storage unit

134: Surface temperature storage unit

1211: Sleep and wakefulness determination department

1212: Sleeping detection department

1213: Wake-up detection department

1214: Mid-course sobriety detection department

1231: Face position detection unit

1232: Surface temperature measurement unit

1233: Average body temperature calculation unit

1234: Abnormal body temperature determination department

S: judgment value

Y1: Arrow

ZCM: Activity volume

S1~S11: Steps

FIG1 is a block diagram showing an example of the structure of a physical condition prediction system according to an embodiment of the present disclosure.

FIG2 is a block diagram showing an example of the configuration of the server shown in FIG1.

FIG3 is a diagram showing the structure of the sleep determination unit shown in FIG2.

FIG4 is a diagram showing the structure of the body temperature determination unit shown in FIG2.

FIG5 is a diagram showing an example of the sleep state output from the sleep determination unit in one day.

FIG6 is a diagram showing an example of the sleep state output from the sleep determination unit within a predetermined period.

Figure 7 is a diagram showing an example of body status data within a predetermined period.

Figure 8 is a flow chart used to illustrate the actions of the server in this embodiment.

Figure 9 is a diagram showing an example of the history of the standard deviation and mean value of body movement data.

Figure 10 is a graph used to illustrate the correlation between the onset of peripheral symptoms of dementia (BPSD) and sleep status.

The form used to implement the invention

(Knowledge and insights that serve as the basis for this disclosure)

In the previous technology, the display unit or speaker installed in the sleep sensor is driven and controlled according to the sleep state of the subject, or the home appliance installed outside the sleep sensor is remotely controlled.

For example, the conventional home appliance control system can open the electric curtains, play wake-up music from the audio equipment, turn on the lighting equipment, turn on the news channel on the TV, set the bedroom temperature to an appropriate level with the air conditioning equipment, and adjust the bedding to a state that is easy for the subject to get up (such as tilt adjustment of the electric bed or pressure adjustment of the air mattress). In this way, in the conventional technology, the sleep sensor and various home appliances are coordinated to provide the subject with a comfortable waking experience.

However, in the prior art, although there are methods for controlling the equipment in response to the sleep state of the subject analyzed by the sleep sensor, there is no consideration of predicting changes in the physical condition of the subject.

In order to solve the above problems, a method for predicting physical condition in one aspect of the present disclosure obtains physiological data of a subject, continuously determines the sleep state of the subject based on the obtained physiological data, and predicts changes in the physical condition of the subject based on the determined sleep state.

Based on this structure, the physiological data of the target person is obtained. Based on the obtained physiological data, the sleep state of the target person is continuously determined. Based on the determined sleep state, the change of the physical condition of the target person is predicted.

Therefore, it is possible to predict changes in the physical condition of the target person based on the continuously determined sleep state. Also, since changes in the physical condition of the target person are predicted, for example, when the target person is an elderly person or a patient with dementia, the care plan for the target person can be re-evaluated, and the target person can be cared for more efficiently.

Furthermore, in the above-mentioned body condition prediction method, the above-mentioned physiological data includes body movement data showing the body movement of the above-mentioned object person, and the above-mentioned determination can also continuously determine the above-mentioned sleep state based on the above-mentioned body movement data.

According to this structure, the physiological data includes body movement data showing the body movement of the target person. During the determination, the sleep state is continuously determined based on the body movement data. Therefore, since the sleep state is continuously determined based on the body movement data showing the body movement of the target person, the sleep state of the target person can be accurately determined.

Furthermore, in the above-mentioned physical condition prediction method, the above-mentioned prediction can also predict the deterioration of the physical condition of the above-mentioned object person when the above-mentioned physical movement data drops below a predetermined value within a predetermined period.

According to this configuration, in the prediction, when the body motion data drops below a predetermined value within a predetermined period, it is predicted that the physical condition of the target person will deteriorate. Therefore, the deterioration of the physical condition of the target person can be accurately predicted by the body motion data.

Furthermore, in the above-mentioned method for predicting the physical condition, physical condition data showing whether the physical condition of the above-mentioned object person is good or not is obtained, and the above-mentioned prediction can also predict the change of the above-mentioned physical condition from the correlation between the above-mentioned sleep state process determined and the above-mentioned physical condition data process obtained.

Based on this structure, the physical condition data showing whether the physical condition of the target person is good or not is obtained. In the prediction, the change of the physical condition is predicted from the correlation between the course of the determined sleep state and the course of the obtained physical condition data.

Therefore, if there is a correlation between the course of sleep state and the course of physical state data, the correlation can be used to easily predict changes in the physical condition of the subject.

Furthermore, in the above-mentioned physical condition prediction method, the body temperature of the aforementioned target person is further detected, and then it is determined whether the aforementioned body temperature of the aforementioned target person is higher than a predetermined temperature. Moreover, when it is determined that the aforementioned body temperature is higher than the predetermined temperature, it is also possible to predict the deterioration of the physical condition of the aforementioned target person.

According to this structure, the body temperature of the target person is detected. It is judged whether the body temperature of the target person is higher than the predetermined temperature. When it is judged that the body temperature is higher than the predetermined temperature, it is predicted that the physical condition of the target person will deteriorate.

Therefore, since it is predicted that the physical condition of the target person will deteriorate when the body temperature of the target person is judged to be higher than a predetermined temperature, changes in the physical condition of the target person can be easily predicted using the body temperature of the target person.

Furthermore, in the above-mentioned method for predicting physical condition, the above-mentioned prediction can also be based on the frequency of the above-mentioned subject being awake at night to predict the deterioration of the above-mentioned subject's physical condition.

According to this configuration, in the prediction, the deterioration of the physical condition of the subject is predicted according to the frequency of the subject waking up at night. When the frequency of the subject waking up at night is high, there is a possibility that the sleep rhythm of the subject has collapsed. Therefore, the deterioration of the physical condition of the subject can be accurately predicted according to the frequency of the subject waking up at night.

Furthermore, in the above-mentioned method for predicting physical condition, the above-mentioned prediction can also predict the deterioration of the physical condition of the above-mentioned subject when the frequency of the above-mentioned subject waking up at night is more than a predetermined number of times.

According to this structure, in the prediction, when the frequency of the subject person waking up at night is more than a predetermined number of times, it is predicted that the physical condition of the subject person will deteriorate, and therefore, the deterioration of the physical condition of the subject person can be accurately predicted.

Furthermore, in the above-mentioned physical condition prediction method, the deterioration of the physical condition of the aforementioned subject includes the onset of peripheral symptoms of dementia, and the above-mentioned prediction can also be based on at least one of the frequency of the aforementioned subject being awake at night and the frequency of the subject taking a nap or sleeping in the evening to predict the onset of peripheral symptoms of dementia of the aforementioned subject.

According to this constitution, the deterioration of the subject's physical condition includes the onset of peripheral symptoms of dementia. In the prediction, the onset of peripheral symptoms of dementia of the subject is predicted based on at least one of the frequency of the subject's awakening at night and the frequency of the subject's nap or evening sleep.

Therefore, the onset of peripheral symptoms of dementia in a subject can be reliably predicted based on at least one of the subject's frequency of being awake at night and the subject's frequency of napping or sleeping in the evening.

Furthermore, in the above-mentioned method for predicting physical condition, the above-mentioned prediction can also predict the deterioration of the physical condition of the above-mentioned subject based on the time that the above-mentioned subject stays awake at night.

According to this configuration, in the prediction, the deterioration of the physical condition of the target person is predicted according to the time the target person stays awake at night. When the target person stays awake for a long time at night, there is a possibility that the sleep rhythm of the target person has collapsed. Therefore, it is possible to accurately predict the deterioration of the physical condition of the target person according to the time the target person stays awake at night.

Furthermore, in the above-mentioned physical condition prediction method, the above-mentioned prediction may also predict the deterioration of the physical condition of the above-mentioned object person when the above-mentioned object person stays awake for more than a predetermined time at night.

According to this configuration, when the subject person stays awake for more than a predetermined time at night, the subject person's physical condition is predicted to deteriorate, and thus the subject person's physical condition can be accurately predicted to deteriorate.

Furthermore, in the above-mentioned physical condition prediction method, the prediction result of the change of the physical condition of the above-mentioned target person can be further sent to the terminal device.

According to this structure, since the prediction result of the change of the physical condition of the target person is sent to the terminal device, the prediction result can be reported to the administrator through the terminal device.

The physical condition prediction device of another aspect of the present disclosure is equipped with a communication unit and a processor. The communication unit obtains the physiological data of the target person. The processor continuously determines the sleep state of the target person based on the obtained physiological data, and predicts the change of the physical condition of the target person based on the determined sleep state.

Based on this structure, the physiological data of the target person is obtained. Based on the obtained physiological data, the sleep state of the target person is continuously determined. Based on the determined sleep state, the change of the physical condition of the target person is predicted.

Therefore, it is possible to predict changes in the physical condition of the target person based on the continuously determined sleep state. Also, since changes in the physical condition of the target person are predicted, for example, when the target person is an elderly person or a patient with dementia, the care plan for the target person can be re-evaluated, and the target person can be cared for more efficiently.

The physical condition prediction program of other aspects of the present disclosure enables the processor to perform the following processing: continuously determine the sleep state of the aforementioned target person based on the physiological data of the aforementioned target person, and predict the change of the physical condition of the aforementioned target person based on the determined aforementioned sleep state.

Based on this structure, the sleeping state of the target person is continuously determined based on the physiological data obtained. Based on the determined sleeping state, the changes in the physical condition of the target person are predicted.

Therefore, it is possible to predict changes in the physical condition of the target person based on the continuously determined sleep state. Also, since changes in the physical condition of the target person are predicted, for example, when the target person is an elderly person or a patient with dementia, the care plan for the target person can be re-evaluated, and the target person can be cared for more efficiently.

The following is a description of the implementation of the present disclosure with reference to the attached drawings. Furthermore, the following implementation is an example of the present disclosure and does not limit the technical scope of the present disclosure.

(Implementation form)

FIG1 is a block diagram showing an example of the structure of a body condition prediction system of an embodiment of the present disclosure. The body condition prediction system shown in FIG1 comprises: a server 1, a motion sensor 2, a temperature sensor 3, and a terminal device 4.

The server 1 is connected to the motion sensor 2, the temperature sensor 3 and the terminal device 4 via the network 5 so as to be able to communicate. Furthermore, the network 5 is, for example, the Internet.

The motion sensor 2 is, for example, a Doppler sensor, which is installed on the ceiling or wall of the target person's room. The target person is, for example, a resident of an elderly-oriented residence, such as an elderly person or a person under care. The motion sensor 2 emits radio waves, and detects the target person's movements by comparing the frequency of the radio waves reflected when hitting the target person with the frequency of the emitted radio waves. The motion sensor 2 continuously detects the body movements of the target person, and continuously sends body movement data showing the detected body movements of the target person to the server 1. It is ideal for the motion sensor 2 to continuously detect the body movements of the target person, for example, at intervals of 1 second. Furthermore, the motion sensor 2 can also continuously detect the body movements of the target person, for example, at intervals of 1 minute, and the detection interval is not particularly limited. In addition, body movement data is an example of physiological data. In addition to the body movements of the target person, the motion sensor 2 can also detect the pulse and breathing of the target person.

Furthermore, the motion sensor 2 may be, for example, an acceleration sensor. In this case, the motion sensor 2 is mounted on the body of the target person to detect the body movements of the target person.

Furthermore, the motion sensor 2 can also be incorporated into household electrical appliances such as lighting devices installed in the room.

The temperature sensor 3 is, for example, an infrared camera, which is installed on the ceiling or wall of the target person's room. The temperature sensor 3 continuously captures infrared images in the room and continuously sends the captured infrared images to the server 1.

Furthermore, the temperature sensor 3 can also be incorporated into household appliances such as air conditioning equipment installed in the room.

The terminal device 4 is, for example, a personal computer or a tablet computer, and is operated by the administrator of the physical condition of the managed object. The terminal device 4 receives the physical condition data input by the administrator of the physical condition data to show whether the physical condition of the object is good. For example, the terminal device 4 receives the input of the physical condition data to show whether the physical condition of the object is good every day. Furthermore, the terminal device 4 can also receive the input of the physical condition data to show whether the physical condition of the object is good at each predetermined time or each predetermined time band. The terminal device 4 sends the input physical condition data to the server 1.

In addition to receiving input of information about whether the physical condition of the target person is good, the terminal device 4 can also receive input of other information. For example, the terminal device 4 can also receive input of information related to nursing records, which are about the type and time of administration of the medicine that has been administered to the target person.

Furthermore, the motion sensor 2 and the temperature sensor 3 can send the sensing data directly to the server 1, or send it to the server 1 through the terminal device 4.

FIG2 is a block diagram showing an example of the configuration of the server of FIG1. The server 1 shown in FIG2 includes: a communication unit 11, a control unit 12, and a memory unit 13.

The communication unit 11 includes: a body movement data acquisition unit 111, a body status data acquisition unit 112, an infrared image acquisition unit 113, and a prediction result sending unit 114.

The body motion data acquisition unit 111 acquires the body motion data of the body motion of the display object person. The body motion data acquisition unit 111 receives the body motion data sent by the motion sensor 2.

The body motion data acquisition unit 111 can also acquire body motion data sent from the body motion sensor. The body motion data acquisition unit 111 can also acquire body motion values once every minute, for example.

The body state data acquisition unit 112 acquires body state data indicating whether the body state of the target person is good or not. The body state data acquisition unit 112 receives the body state data sent by the terminal device 4.

The body condition data acquisition unit 112 can also acquire recorded information of the caregiver, such as nursing records. The body condition information data includes, for example, vital signs such as body temperature or blood pressure, observation records made subjectively by the caregiver, the presence or absence of falls, and BPSD (wandering, delusions, etc.).

The infrared image acquisition unit 113 acquires an infrared image. The infrared image acquisition unit 113 receives the infrared image sent by the temperature sensor 3.

The control unit 12 is, for example, a CPU (central processing unit) and controls the entire server 1. The control unit 12 includes: a sleep determination unit 121, a body condition analysis unit 122, a body temperature determination unit 123, and a body condition prediction unit 124.

The memory unit 13 is, for example, a semiconductor memory or a hard disk, and includes: a sleep state storage unit 131, a body state data storage unit 132, and a body state prediction information storage unit 133.

The sleep determination unit 121 continuously determines the sleep state of the target person based on the body movement data acquired by the body movement data acquisition unit 111.

The sleep determination unit 121 can also determine sleep/wakefulness based on the 7-minute body movements obtained by the body movement data acquisition unit 111. Sleep/wakefulness is determined based on the Cole algorithm also used in medical equipment (such as activity recorders). In addition, the acquired body movement data is accompanied by absolute time information.

The sleep state storage unit 131 stores the sleep state history of the target person determined by the sleep determination unit 121. The sleep state storage unit 131 stores the state of the target person being asleep or awake in a predetermined time unit. The predetermined time unit is, for example, 1 minute or 1 second.

The sleep state storage unit 131 can also store the sleep/wakefulness determined by the sleep determination unit 121 together with the time information.

FIG3 is a diagram showing the structure of the sleep determination unit shown in FIG2. The sleep determination unit 121 shown in FIG3 includes: a sleep wake determination unit 1211, a sleep detection unit 1212, a wake-up detection unit 1213, and a mid-wake detection unit 1214. The body movement data includes the activity volume (size of movement) ZCM per minute, and is input from the body movement data acquisition unit 111 to the sleep wake determination unit 1211.

The sleep wakefulness determination unit 1211 uses the following formula (1) to calculate the determination value S. In the following formula (1), ZCM _-4min indicates the amount of activity 4 minutes ago, ZCM _-3min indicates the amount of activity 3 minutes ago, ZCM _-2min indicates the amount of activity 2 minutes ago, ZCM _-1min indicates the amount of activity 1 minute ago, ZCM _now indicates the amount of activity at the determination time, ZCM _+1min indicates the amount of activity 1 minute later, and ZCM _+2min indicates the amount of activity 2 minutes later.

S=0.0033(1.06ZCM _-4min +0.54ZCM _-3min +0.58ZCM _-2min +0.76ZCM _-1min +2.3ZCM _now +0.74ZCM _+1min +0.67ZCM _+2min )‧‧‧(1)

When the judgment value S is greater than 1, the sleep-wake judgment unit 1211 judges that the target person is awake, and when the judgment value S is less than 1, it judges that the target person is sleeping.

The sleep detection unit 1212 detects the first time when the subject sleeps continuously for a predetermined time or longer as the sleep time when the subject falls asleep. The sleep detection unit 1212 outputs the detected sleep time to the sleep state storage unit 131 and the body condition prediction unit 124.

The waking up detection unit 1213 detects the first time when the subject is awake for a predetermined time or longer as the waking up time. The waking up detection unit 1213 outputs the detected waking up time to the sleep state storage unit 131 and the body condition prediction unit 124.

The mid-awakening detection unit 1214 detects and determines the time of continuous awakening from the time of falling asleep to the time of waking up as the mid-awakening time when the subject wakes up during sleep. The mid-awakening detection unit 1214 outputs the detected mid-awakening time to the sleep state storage unit 131 and the body condition prediction unit 124.

Furthermore, the sleep determination unit 121 can output any determination result of sleep and wakefulness determined by the sleep wakefulness determination unit 1211 as a sleep state to the sleep state storage unit 131 in addition to the sleep time, wakefulness time and wakefulness time. The sleep state storage unit 131 can store the time change of whether the subject is sleeping or awake as the sleep state in addition to the sleep time, wakefulness time and wakefulness time.

Furthermore, although in this embodiment, the sleep-wake determination unit 1211 uses the above-mentioned formula (1) to calculate the determination value S, the present disclosure is not particularly limited thereto, and other formulas such as the following formula (2) may also be used to calculate the determination value S.

S=0.00001(404ZCM _-4min +598ZCM _-3min +326CM _-2min +441ZCM _-1min +1408ZCM _now +508ZCM _+1min +350ZCM _+2min )‧‧‧(2)

The above formula (2) is also called Cole's algorithm, which is a general formula used for sleep determination (Roger J. Cole, Daniel F. Kripke, William Gruen, Daniel J. Mullaney, J. Christian Gillin, "Automatic Sleep/Wake Identification From Wrist Activity", 15(5), 461-469, 1992). In the following formula (2), ZCM _-4min shows the activity amount 4 minutes ago, ZCM _-3min shows the activity amount 3 minutes ago, ZCM _-2min shows the activity amount 2 minutes ago, ZCM _-1min shows the activity amount 1 minute ago, ZCM _now shows the activity amount at the determination time, ZCM _+1min shows the activity amount 1 minute later, and ZCM _+2min shows the activity amount 2 minutes later.

The body state data storage unit 132 stores the body state data acquired by the body state data acquisition unit 112. Similar to the sleep state storage unit 131, the body state data storage unit 132 can also store the body state data together with the time information of the records related to the body state.

The body condition analysis unit 122 analyzes the tendency of the subject's sleep state before the subject's physical condition deteriorates based on the correlation between the body condition data indicating whether the subject's physical condition is good during the predetermined period and the subject's sleep state during the predetermined period.

For example, when the body state analysis unit 122 reads the body state data of the predetermined period from the body state data storage unit 132, it also reads the sleep state of the predetermined period from the sleep state storage unit 131. The predetermined period is, for example, one month. The body state analysis unit 122 analyzes the body state data and the sleep state within the predetermined period, and if the lack of sleep lasts for 2 days before the body condition deteriorates, the body condition prediction information that the body condition will deteriorate after the lack of sleep continues for 2 days is generated. Furthermore, for example, when the sleep time between 19:00 and 7:00 does not reach the predetermined time, it is determined that the sleep is insufficient. The body condition analysis unit 122 stores the generated body condition prediction information in the body condition prediction information storage unit 133.

For example, the body state analysis unit 122 analyzes body state data and sleep state for one month, and if the frequency of waking up during sleep from the time of falling asleep to the time of waking up reaches a predetermined number of times or more on the day before the body state deteriorates, body state prediction information is generated that the body state will deteriorate the next day if the frequency of waking up during sleep reaches a predetermined number of times or more.

In addition, for example, the body state analysis unit 122 analyzes the body state data and sleep state for one month, and if the total time of being awake during sleep from the time of falling asleep to the time of waking up on the day before the body condition deteriorates reaches a predetermined time or more, the body condition prediction information that the body condition will deteriorate the next day is generated.

The physical condition analysis unit 122 can also learn the prediction model by inputting the physical condition data indicating whether the physical condition of the target person during the predetermined period is good or not and the sleep state of the target person during the predetermined period as teaching data into the prediction model for predicting the deterioration of the physical condition of the target person, and store the prediction model as physical condition prediction information in the physical condition prediction information storage unit 133.

The body state analysis unit 122 is a part that cross-analyzes the sleep state from the sleep state storage unit 131 and the body condition data from the body state data storage unit 132. The body state analysis unit 122 can also capture the emergency information related to the body condition stored in the body state data storage unit 132. The so-called emergency information refers to, for example, a fall and the time, a fever and the time, a BPSD attack and the time. In addition, the body state analysis unit 122 can also capture the sleep state just before the time when the emergency occurs when the emergency has occurred. In this way, the body condition analysis unit 122 can also store the sleep state or body temperature change state before the emergency event in the body condition prediction information storage unit 133 according to the state of the elderly and the location of each emergency event information to be captured.

The body condition prediction information storage unit 133 stores body condition prediction information used to predict changes in the body condition of the target person. Furthermore, the correlation between the body condition data and the sleep state varies for each target person. Therefore, the body condition prediction information is information inherent to the target person, stored in the body condition prediction information storage unit 133 and set to be related to the target person.

The body temperature determination unit 123 determines whether the body temperature of the target person is higher than a predetermined temperature based on the infrared image acquired by the infrared image acquisition unit 113.

FIG4 is a diagram showing the structure of the body temperature determination unit shown in FIG2. The body temperature determination unit 123 shown in FIG4 includes: a face position detection unit 1231, a surface temperature measurement unit 1232, an average body temperature calculation unit 1233, and an abnormal body temperature determination unit 1234. The memory unit 13 includes a surface temperature storage unit 134.

The face position detection unit 1231 detects the face position of the target person from the infrared image acquired by the infrared image acquisition unit 113. The face position detection unit 1231 detects the face position of the target person from the infrared image, for example, by pattern matching.

The surface temperature measuring unit 1232 measures the surface temperature of the face position detected by the face position detecting unit 1231.

The surface temperature storage unit 134 stores the surface temperature of the face position measured by the surface temperature measuring unit 1232.

The average body temperature calculation unit 1233 calculates the average value of the surface temperatures at the face position stored in the surface temperature storage unit 134 as the average body temperature.

The abnormal body temperature determination unit 1234 determines whether the surface temperature of the face position measured by the surface temperature measurement unit 1232 is higher than the average body temperature calculated by the average body temperature calculation unit 1233. When it is determined that the measured surface temperature of the face position is higher than the average body temperature, the abnormal body temperature determination unit 1234 determines that the body temperature of the subject person is abnormal. In addition, when it is determined that the measured surface temperature of the face position is below the average body temperature, the abnormal body temperature determination unit 1234 determines that the body temperature of the subject person is normal.

The body condition prediction unit 124 predicts changes in the body condition of the target person based on the sleep state determined by the sleep determination unit 121. The body condition prediction unit 124 predicts changes in the body condition based on the correlation between the history of the sleep state determined by the sleep determination unit 121 and the history of the body condition data acquired by the body condition data acquisition unit 112.

The body condition prediction unit 124 refers to the body condition prediction information, which is stored in the body condition prediction information storage unit 133 and is produced by the correlation between the sleep state within the predetermined period and the body condition data within the predetermined period. When the sleep state determined by the sleep determination unit 121 meets the conditions specified by the body condition prediction information, the body condition prediction unit 124 predicts that the body condition will deteriorate. For example, when the total sleep time at night is less than the predetermined time for two consecutive days, the body condition prediction unit 124 predicts that the body condition of the subject will deteriorate the next day.

Furthermore, the physical condition prediction unit 124 can also predict the deterioration of the physical condition of the target person according to the frequency of the target person waking up at night. That is, the physical condition prediction unit 124 can also predict the deterioration of the physical condition of the target person when the frequency of the target person waking up at night is more than a predetermined number of times. For example, the physical condition prediction unit 124 can also predict that the physical condition of the target person will deteriorate the next day when the frequency of waking up during sleep at night reaches more than a predetermined number of times.

In addition, the physical condition prediction unit 124 can also predict the deterioration of the physical condition of the target person according to the time the target person stays awake at night. That is, the physical condition prediction unit 124 can also predict the deterioration of the physical condition of the target person when the time the target person stays awake at night is more than a predetermined time. For example, the physical condition prediction unit 124 can also predict that the physical condition of the target person will deteriorate the next day when the total time of waking up during sleep at night reaches more than a predetermined time.

Furthermore, the physical condition prediction unit 124 can also predict the deterioration of the physical condition of the target person when the body temperature determination unit 123 determines that the body temperature of the target person is higher than a predetermined temperature.

In addition, the body condition prediction unit 124 can also perform a similarity analysis (pattern matching) on the sleep state input from the sleep determination unit 121 and the sleep state when the emergency occurs stored in the body condition prediction information storage unit 133 in real time. The body condition prediction unit 124 calculates the similarity between the input sleep pattern and the sleep pattern when the emergency occurs based on the result of the similarity analysis, and can predict poor physical condition from the sleep data based on the threshold value of the similarity.

In this way, the system uses the physical condition data acquisition unit 112 to learn the sleep or body temperature change patterns associated with poor physical conditions as typical examples as teaching data, thereby alerting the elderly of unexpected events that the caregiver cannot fully predict.

Furthermore, in this embodiment, although an example is shown in which the physical condition data acquisition unit 112 acquires the physical condition data of the person in the past who is predicted to be in a poor physical condition, the present invention is not limited thereto. The physical condition data acquisition unit 112 may also acquire the physical condition data of others who have the same medical history or need for care, and the physical condition prediction unit 124 may also predict the physical condition of the person based on the physical condition data of others.

In addition, in this embodiment, although an example of the body condition prediction unit 124 performing a likelihood analysis is shown, it is not limited to this. For example, in order to detect abnormalities in vital signs, the same effect can be obtained by using accuracy analysis such as a hidden Markov model. By using accuracy analysis, even if there are no observation records of caregivers such as nursing records, it is possible to find abnormal values of vital signs, set the situation where the abnormal value occurs as an emergency, and calculate (predict) the occurrence prediction of the same emergency from the sleep state.

Furthermore, in this embodiment, although the physical condition is predicted from sleep/body temperature, it is not limited to this. The sleep information of this system is originally calculated from the body movements of the elderly, that is, the activity status. Therefore, the physical condition can also be predicted from the pattern analysis between the activity status of the elderly and sudden events. For example, it can predict the reduction in activity caused by poor physical condition or dehydration.

In terms of vital signs, in addition to body temperature, blood pressure can also achieve the same effect as above. In particular, regarding blood pressure, wrist-type blood pressure monitors have recently been commercialized. If you use this wrist-type blood pressure monitor, you can obtain continuous blood pressure data and analyze the pattern between blood pressure changes and sudden events.

In addition, in the likelihood analysis of the physical condition prediction unit 124, correlation analysis in deep learning or machine learning can be performed. In particular, for elderly people, the emergency information related to the physical condition stored in the physical condition data storage unit 132 is mostly complex information. It is necessary to include multiple correlations to analyze how such complex emergency information is related to the life rhythm such as sleep. In addition, in addition to the correlation between the data of the person himself, when the correlation between the data of others is also considered, the correlation analysis will become complicated. In such a case, correlation analysis can be performed by deep learning or machine learning.

The prediction result sending unit 114 sends the physical condition prediction result that predicts the change of the physical condition of the target person to the terminal device 4. When the prediction result sending unit 114 predicts the deterioration of the physical condition of the target person, it sends the physical condition prediction result to the terminal device 4.

The terminal device 4 receives the body condition prediction result sent from the server 1, and notifies the administrator of the received body condition prediction result. The terminal device 4, for example, displays the received body condition prediction result. In addition, the terminal device 4 can also output the received body condition prediction result by sound, for example. Furthermore, the terminal device 4 that notifies the body condition prediction result can be the same as the terminal device that receives the input of the body condition data, or it can be different.

FIG5 is a diagram showing an example of the sleep status output from the sleep determination unit in one day. FIG5 shows the sleep status of the subject person in one day from 7:00 a.m. to 6:59 a.m. the next day. In this embodiment, the motion sensor 2 is arranged in the subject person's room, and sleep is determined based on the body motion data detected by the motion sensor 2. Therefore, the sleep determination unit 121 can determine whether the subject person is in the room in addition to whether the subject person is sleeping or awake. On the contrary, when the subject person is not in the room, the sleep determination unit 121 cannot determine whether the subject person is sleeping or awake.

In FIG5 , the horizontal axis represents time, and the vertical axis represents whether the subject is sleeping, the subject is indoors (awake), or the subject is not indoors. The subject's sleeping, indoors (awake), and absence are represented by bar graphs. When the subject is not indoors, the level of the bar graph is 0 (the bottom of FIG5 ); when the subject is indoors and awake, the level of the bar graph is 1 (the middle of FIG5 ); when the subject is indoors and sleeping, the level of the bar graph is 2 (the top of FIG5 ). The bar graph is displayed in units of 1 minute, for example.

FIG6 is a diagram showing an example of the sleep state output from the sleep determination unit during a predetermined period. FIG6 shows the sleep state of the subject person from September 7 to October 2 of a certain year. As shown in FIG6, the subject person did not get enough sleep on the nights of September 11, 12, 14, 17, 18, 20, 22, 23, 25, 26, 27, 29, 30, and October 2.

FIG. 7 is a diagram showing an example of physical condition data within a predetermined period. FIG. 7 shows the physical condition of the subject from September 7 to October 2 of a certain year. In FIG. 7, ○ indicates that the physical condition of the subject is good, and × indicates that the physical condition of the subject is poor. The sleep state shown in FIG. 6 and the physical state shown in FIG. 7 are data showing the same subject. There is a correlation between the sleep state shown in FIG. 6 and the physical state shown in FIG. 7 that the days of insufficient sleep lasted for 2 days before the physical condition deteriorated. For example, the subject was insufficiently slept for 2 consecutive days on September 11 and 12, and the physical condition deteriorated on the next day, September 13.

In this way, the body condition analysis unit 122 analyzes the body condition data and sleep status within a predetermined period, and generates body condition prediction information that the body condition will deteriorate the next day after insufficient sleep continues for two days before the body condition deteriorates.

Figure 8 is a flow chart used to illustrate the actions of the server in this embodiment.

First, in step S1, the physical condition data acquisition unit 112 acquires physical condition data indicating whether the physical condition of the target person is good. The physical condition data acquisition unit 112 receives the physical condition data sent by the terminal device 4. The physical condition data, for example, indicates whether the physical condition of the target person was good the day before. Here, if it is determined that the physical condition data has not been acquired (No in step S1), the processing moves to step S6.

On the other hand, if it is determined that the body state data has been obtained (Yes in step S1), in step S2, the body state data acquisition unit 112 stores the obtained body state data in the body state data storage unit 132. Furthermore, the body state data acquisition unit 112 can obtain body state data showing the body state of one day, and can also obtain body state data showing the body state of multiple days.

Next, in step S3, the body state analysis unit 122 determines whether a predetermined period of time has passed since the body state data of the target person was acquired. For example, the body state analysis unit 122 determines whether one month has passed since the body state data of the target person was acquired. Furthermore, the predetermined period of time is not limited to one month.

Here, if it is determined that the predetermined time has not passed since the acquisition of the target person's physical status data began (No in step S3), the process moves to step S6.

On the other hand, if it is determined that the predetermined time has passed since the acquisition of the target person's physical state data began (Yes in step S3), in step S4, the physical state analysis unit 122 analyzes the correlation between the physical state of the target person within the predetermined period and the sleep state of the target person within the predetermined period. The physical state analysis unit 122 generates physical state prediction information for predicting changes in the physical state of the target person based on the analysis results.

Next, in step S5, the body condition analysis unit 122 stores the generated body condition prediction information in the body condition prediction information storage unit 133.

Furthermore, when the body state and the sleep state are related, the change of the body state can be predicted by the sleep state alone. Therefore, when the body state prediction information storage unit 133 has stored the body state prediction information, the body state data acquisition unit 112 can also stop acquiring the body state data. Moreover, the body state data acquisition unit 112 can also stop acquiring the body state data when the body state prediction information storage unit 133 does not store the body state prediction information, without considering the relationship between the body state and the sleep state.

In addition, the body state data acquisition unit 112 can also restart the acquisition of body state data when the body state prediction information storage unit 133 does not store the body state prediction information, and the body state analysis unit 122 can extend the expected period without considering the correlation between the body state and the sleep state. By extending the expected period, the possibility of discovering the correlation between the body state and the sleep state can be increased.

Next, in step S6, the body motion data acquisition unit 111 acquires the body motion data of the display object character.

Next, in step S7, the sleep determination unit 121 determines whether the display object person is sleeping or awake based on the body motion data acquired by the body motion data acquisition unit 111.

Next, in step S8, the sleep determination unit 121 stores the determined sleep state of the target person in the sleep state storage unit 131.

Next, in step S9, the body condition prediction unit 124 predicts changes in the body condition of the target person based on the sleep state determined by the sleep determination unit 121 and the body condition prediction information stored in the body condition prediction information storage unit 133.

Next, in step S10, the physical condition prediction unit 124 determines whether it is predicted that the physical condition of the target person will deteriorate. Here, if it is determined that the physical condition of the target person is not predicted to deteriorate (No in step S10), the process returns to step S1.

On the other hand, when it is determined that the physical condition of the target person is predicted to deteriorate (yes in step S10), in step S11, the prediction result sending unit 114 sends the physical condition prediction result predicting the deterioration of the physical condition of the target person to the terminal device 4. Then, the processing returns to step S1.

In this way, in the physical condition prediction system of this embodiment, it is possible to predict changes in the physical condition of the target person based on the continuously determined sleep state. In addition, since changes in the physical condition of the target person are predicted, for example, when the target person is an elderly person or a patient with dementia, the care plan for the target person can be re-evaluated, and the target person can be cared for more efficiently.

Furthermore, in this embodiment, although the body condition prediction unit 124 predicts the change of the body condition of the target person at the time point of determining the sleep state, this disclosure is not particularly limited to this, and the body condition prediction unit 124 can also predict the change of the body condition of the target person at a predetermined time point. The predetermined time point is, for example, a predetermined time such as 7 a.m. every day, or a predetermined time such as every hour. At this time, after the processing of step S8, the body condition prediction unit 124 determines whether it is a predetermined time point. And, when it is determined to be a predetermined time point, the processing is transferred to step S9, and when it is determined to be not a predetermined time point, the processing can also return to step S1.

Furthermore, in this embodiment, the physical condition prediction unit 124 can also predict the deterioration of the physical condition of the target person when the physical movement data drops below a predetermined value within a predetermined period.

FIG9 is a diagram showing an example of the history of the standard deviation and average value of body motion data. In FIG9, the standard deviation and average value of the body motion data of one day are displayed every five days. As shown by arrow Y1 in FIG9, when the standard deviation of the body motion data decreases sharply during a predetermined period, there is a possibility that the daily activities (ADL) of the subject person decreases and the physical condition of the subject person deteriorates. Therefore, the physical condition prediction unit 124 can also predict the deterioration of the physical condition of the subject person when the body motion data decreases below a predetermined value during a predetermined period.

Figure 10 is a graph used to illustrate the correlation between the onset of peripheral symptoms of dementia (BPSD) and sleep status. In Figure 10, the number of naps, the number of awake times at night, the awake time at night, and the nighttime actions of the subject are displayed daily.

The sleep state is collected in the sleep state storage unit 131 through the sleep determination unit 121. In addition, the body condition data acquisition unit 112 extracts data that matches the symptoms of BPSD such as wandering or delusions from the nursing record data, and stores it in the body condition data storage unit 132. The nursing record data is input by the caregiver of the target person, for example.

According to FIG. 10 , it is found that there are records of BPSD attacks of wandering on May 21 and delusions on May 22 and May 27. The body state analysis unit 122 analyzes the main causes of BPSD attacks based on BPSD attacks and data stored in the sleep state storage unit 131. In this example, the following correlation is derived: the number of awakenings in the middle of the night is high on days when BPSD is confirmed, and the number of naps or evening sleeps is high on days when the number of awakenings in the middle of the night is high. Therefore, it is predicted that naps or evening sleeps that cannot be managed by caregivers are related to BPSD attacks at night.

Therefore, the body condition analysis unit 122 analyzes the correlation between the frequency of the subject's awakeness at night, the frequency of the subject's nap or evening sleep, and the onset of the subject's peripheral symptoms of dementia, and generates body condition prediction information for predicting the onset of the subject's peripheral symptoms of dementia based on the frequency of the subject's awakeness at night and the frequency of the subject's nap or evening sleep. The body condition prediction unit 124 predicts the onset of the subject's peripheral symptoms of dementia based on the frequency of the subject's awakeness at night and the frequency of the subject's nap or evening sleep.

Furthermore, the physical condition prediction unit 124 can also predict the onset of peripheral symptoms of dementia of the subject person according to the frequency of the subject person's awakeness at night. Furthermore, the physical condition prediction unit 124 can also predict the onset of peripheral symptoms of dementia of the subject person according to the frequency of the subject person's nap or evening sleep. The physical condition prediction unit 124 can also predict the onset of peripheral symptoms of dementia of the subject person according to at least one of the frequency of the subject person's awakeness at night and the frequency of the subject person's nap or evening sleep.

In this way, if the body condition analysis unit 122 stores and analyzes BPSD symptoms and sleep status, the body condition prediction unit 124 can predict the possibility of BPSD at night according to the state of nap or evening sleep. If the onset of BPSD can be predicted, it will help reduce the work burden of the caregiver by preparing for the treatment of BPSD onset in advance, and even eliminate the phenomenon that becomes the main cause of BPSD onset.

In this example, although the example of predicting the onset of BPSD only from the sleep state is described, the present disclosure is not limited to this. The activity amount of the elderly can also be determined from the body movement data as the basis for the analysis of the sleep state, and the correlation between the activity amount and BPSD can be analyzed. In addition, it is also effective to combine various well-known vital sign sensors. The body condition prediction unit 124 can predict the onset of BPSD based on the correlation between the body temperature change obtained by the infrared sensor or the temperature sensor and BPSD. In addition, the body condition prediction unit 124 can predict the onset of BPSD based on the correlation between the heart rate or the breathing rate and BPSD by grasping the heart rate or the breathing rate from the body movement data.

In particular, in terms of heartbeat, the technology of estimating the balance of the autonomic nervous system from changes in heartbeat is already well known. Generally, it is known that excessive stress causes BPSD, and if the degree of stress is confirmed from the heartbeat, the prediction accuracy of BPSD can be further improved.

Furthermore, for example, by combining living environment data such as indoor temperature, indoor humidity, illumination, noise, and carbon dioxide concentration, it is also possible to discover the living environment that may be the main cause of changes in physical condition or the onset of BPSD. Generally speaking, it is known that indoor temperature and indoor humidity will affect the deep body temperature changes that dominate sleep. Therefore, factors that interfere with sleep can also be derived from indoor temperature and indoor humidity. The physical condition prediction unit 124 can predict changes in physical condition or the onset of BPSD based on the correlation between indoor temperature and indoor humidity and changes in physical condition or BPSD. Similarly, factors that interfere with sleep can also be derived from noise or carbon dioxide concentration. The physical condition prediction unit 124 can predict physical condition changes or the onset of BPSD based on the correlation between noise or carbon dioxide concentration and physical condition changes or BPSD.

Although the device disclosed in the present invention is described above based on the implementation form, the present invention is not limited to this implementation form. As long as it does not deviate from the main purpose of the present invention, various modifications that can be imagined by technical personnel in this field are applied to the present implementation form, or the form constructed by combining the constituent elements in different implementation forms can be included in the scope of one or more aspects of the present invention.

Furthermore, in each of the above-mentioned embodiments, each component can be formed by dedicated hardware or can be implemented by executing a software program suitable for each component. Each component can also be implemented by a program execution unit such as a CPU or a processor reading and executing a software program recorded on a recording medium such as a hard disk or a semiconductor memory.

Part or all of the functions of the device of the embodiment of the present disclosure are typically implemented by making an LSI (Large Scale Integration) of an integrated circuit. These functions can be made into a single chip individually, or part or all of the functions can be included in a single chip. In addition, integrated circuitization is not limited to LSI, and can also be implemented using a dedicated circuit or a general-purpose processor. It can also be used: a programmable FPGA (Field Programmable Gate Array) after LSI manufacturing, or a reconfigurable processor that can reconfigure the connection or setting of the circuit battery inside the LSI.

Furthermore, part or all of the functions of the device of the embodiment of the present disclosure may also be implemented by executing a program on a processor such as a CPU.

In addition, all the numbers used above are exemplified for the purpose of specifically explaining the present disclosure, and the present disclosure is not limited to the exemplified numbers.

Furthermore, the execution order of each step shown in the above flowchart is used to specifically illustrate the example of the present disclosure, and other orders other than the above may be used within the scope of obtaining the same result. Furthermore, part of the above steps may also be executed simultaneously (in parallel) with other steps.

In addition, as long as they do not deviate from the main purpose of this disclosure, various modifications of the embodiments of this disclosure that are implemented within the scope that can be imagined by technical personnel in this field are also included in this disclosure.

《Supplementary instructions》

One aspect of the method disclosed herein is: (A) continuously or intermittently obtaining body motion data showing body motions of a subject person; (B) generating sleep state data showing the sleep state of the subject person based on the body motion data; (C) storing the sleep state data in a sleep state database; (D) predicting future changes in the body state of the subject person from the sleep state data with reference to a body state prediction information database; and (E) executing the following (e1) to (e3) when the body state data showing the body state of the subject person has been obtained. (e1) reading past sleep state data within a past period of time from the aforementioned sleep state database; (e2) comparing the aforementioned body state data with the aforementioned past sleep state data, and generating body state prediction information for predicting a specific change of the aforementioned body state from the specific sleep state data; and (e3) registering the aforementioned body state prediction information in the aforementioned body state prediction information database.

For example, the body movement data may be data that is detected by a body movement sensor and displays the number of body movements of the aforementioned target person per unit time. The sleep state data may also be data that displays whether the aforementioned target person is sleeping or staying awake during each period. In the aforementioned (B), (b1) when the evaluation value calculated from the aforementioned body movement data of a plurality of consecutive unit times is smaller than a predetermined value, it is determined that the aforementioned target person is sleeping, and (b2) when the aforementioned evaluation value is greater than the aforementioned predetermined value, it may also be determined that the aforementioned target person is staying awake.

For example, the aforementioned physical state data may also be data input from a terminal of an observer observing the aforementioned object person.

For example, in the aforementioned (e2), it is determined whether there is a correlation between the aforementioned body state data and the aforementioned past sleep state data. When there is the aforementioned correlation, the aforementioned body state data and the aforementioned past sleep state data can be linked to each other to generate the aforementioned body state prediction information.

For example, in the aforementioned (D), the obtained sleep state data can also be pattern matched with the aforementioned past sleep state data in the aforementioned body condition prediction information database.

For example, the body temperature data showing the body temperature of the aforementioned target person may also be obtained continuously or intermittently. In the aforementioned (D), the aforementioned future changes may also be predicted from the aforementioned sleep state data and the aforementioned body temperature data.

For example, in the above (D), the future changes of the above physical state can also be predicted based on the frequency of the above subject's awakening at night.

For example, in the aforementioned (D), the future changes of the aforementioned physical state can also be predicted based on the time the aforementioned subject stays awake at night.

For example, the aforementioned physical condition data may also include information related to the peripheral symptoms of dementia of the aforementioned target person.

For example, the predicted result in (D) above may also be sent to the terminal.

A computer according to one aspect of the present disclosure has a processor and a memory, and the memory records a program useful for enabling the processor to execute any of the above methods.

A non-transitory storage medium according to one aspect of the present disclosure records a program useful for enabling the aforementioned processor to implement any of the above methods.

Industrial availability

The physical condition prediction method, physical condition prediction device, and physical condition prediction program disclosed herein can predict changes in the physical condition of a subject, and are useful as a physical condition prediction method, physical condition prediction device, and physical condition prediction program for predicting the physical condition of a subject.

S1~S11: Steps

Claims (10)

A method for predicting a body condition is a method for predicting a body condition that is executed by a computer having a body motion data acquisition unit, a body state data acquisition unit, a sleep determination unit, and a body state analysis unit, wherein: the body condition prediction method comprises the following steps: the body motion data acquisition unit acquires body motion data showing the body motion of a subject from a motion sensor, the body state data acquisition unit acquires body state data showing whether the body condition of the subject is good or bad from a terminal device, the sleep determination unit continuously determines the sleep state of the subject based on the acquired body motion data, and the sleep determination unit determines the sleep state of the subject based on the acquired body motion data. The sleep state data of the continuously determined sleep state is stored in the sleep state database, the acquired body state data is stored in the body state database, the body state analysis unit compares the body state data showing the poor body condition of the target person read from the body state database with the past sleep state data in a predetermined period before the poor body condition read from the sleep state database to generate body state prediction information, the body state prediction information is used to predict the change of the body condition of the target person from the tendency of the sleep state before the body condition of the target person deteriorates. As in the method for predicting physical condition of claim 1, the aforementioned prediction is to predict the deterioration of the physical condition of the aforementioned object person when the standard deviation of the aforementioned physical motion data drops below a predetermined value within a predetermined period. As in claim 1, the computer has a body temperature determination unit and a body condition prediction unit, and the body condition prediction method includes the following steps: the body temperature determination unit detects the body temperature of the target person, and the body temperature determination unit determines whether the body temperature of the target person is higher than a predetermined temperature, and when the body condition prediction unit determines that the body temperature is higher than the predetermined temperature, it predicts the deterioration of the body condition of the target person. As in the method for predicting physical condition of claim 1, the prediction is made that the physical condition of the subject will deteriorate when the frequency of the subject waking up at night is more than a predetermined number of times. As in claim 1, the physical condition prediction method, wherein the deterioration of the physical condition of the aforementioned subject includes the onset of peripheral symptoms of dementia, and the aforementioned prediction is based on the frequency of the aforementioned subject waking up at night being more than a predetermined number of times, and at least one of the number of naps or evening sleeps of the aforementioned subject, to predict the onset of peripheral symptoms of dementia of the aforementioned subject. As in the method for predicting physical condition of claim 1, the prediction is made that the physical condition of the subject will deteriorate when the subject remains awake for more than a predetermined time at night. As in the method for predicting physical condition of claim 1, the aforementioned computer further performs the following steps: sending the prediction result of the change of the physical condition of the aforementioned object person to the terminal device. As in claim 1, the body condition prediction method, wherein the aforementioned computer has a body condition prediction unit, and the aforementioned body condition prediction method further comprises the following steps: the aforementioned body motion data acquisition unit acquires the first body motion data showing the body motion of the aforementioned target person from the aforementioned motion sensor, the aforementioned sleep determination unit determines the first sleep state of the aforementioned target person based on the acquired first body motion data, and the aforementioned body condition prediction unit predicts the change of the body condition of the aforementioned target person based on the aforementioned first sleep state and the aforementioned body condition prediction information. A computer is provided with a communication unit and a processor, wherein the communication unit has a body motion data acquisition unit and a body state data acquisition unit, and the processor has a sleep determination unit and a body state analysis unit, wherein: the body motion data acquisition unit acquires body motion data showing the body motion of a subject from a motion sensor, the body state data acquisition unit acquires body state data showing whether the body state of the subject is good or bad from a terminal device, the sleep determination unit continuously determines the sleep state of the subject based on the acquired body motion data, and converts the continuously determined sleep state data into sleep data. The sleep state data of the target person is stored in the sleep state database, the body state data obtained is stored in the body state database, the body state analysis unit reads the body state data showing that the body condition of the target person is not good from the body state database, and compares it with the past sleep state data in a predetermined period before the body condition is not good read from the sleep state database, to generate body condition prediction information, the body condition prediction information is used to predict the change of the body condition of the target person from the tendency of the sleep state before the body condition of the target person deteriorates. A body condition prediction program is a program for a computer processor to execute processing. The processor has a sleep determination unit and a body condition analysis unit. The sleep determination unit continuously determines the sleep state of the subject person based on body motion data showing the body motion of the subject person, stores the sleep state data of the continuously determined sleep state in a sleep state database, and stores the body state data showing whether the body condition of the subject person is good or bad in a body condition database. The body state analysis unit compares the body state data showing the poor body state of the target person read from the body state database with the past sleep state data in a predetermined period before the poor body state read from the sleep state database to generate body state prediction information. The body state prediction information is used to predict the change of the body state of the target person from the tendency of the sleep state before the deterioration of the body state of the target person.