Patient closed-loop management system and method based on drug-treatment scheme risk stratification

Abstract

The invention provides a patient closed-loop management system and method based on drug-treatment scheme risk stratification, and relates to the technical field of medical monitoring, and the system comprises a monitoring data module which is used for obtaining medical environment data, body movement data, sleep data and core body health data through an intelligent medical terminal and a sensor; the user data module is used for acquiring historical medical data, basic data and historical monitoring data of the user; the sleep health module is used for determining a sleep health coefficient according to the user historical monitoring data, the user basic data, the historical medical data and the sleep data; the body health module is used for determining a body health coefficient according to the body movement data and the core body health data; the environment safety module is used for determining a medical environment safety coefficient according to the medical environment data; and the monitoring report module is used for determining a monitoring report. According to the invention, comprehensiveness and accuracy of patient closed-loop management can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical monitoring, in particular to a patient closed-loop management system and method based on drug-treatment scheme risk stratification. BACKGROUND

[0002] In the related art, patient closed-loop management mainly relies on sensor data collection combined with manual monitoring, that is, mainly depends on human factors, and excessive reliance on human factors may not guarantee the accuracy and timeliness of management.

[0003] The information disclosed in the background section of this application is only intended to deepen the understanding of the general background of the present application, and should not be regarded as acknowledging or implying in any form that this information constitutes prior art known to those skilled in the art. SUMMARY

[0004] The present application provides a patient closed-loop management system and method based on drug-treatment scheme risk stratification, which can solve the technical problem that the related art cannot guarantee the accuracy and timeliness of management.

[0005] According to a first aspect of the present application, a patient closed-loop management system based on drug-treatment scheme risk stratification is provided, comprising: a monitoring data module for acquiring medical environment data, body movement data, sleep data and core body health data through an intelligent medical terminal and a sensor at multiple time points of a monitoring period; a user data module for acquiring user historical medical data, user basic data and user historical monitoring data; a sleep health module for determining a sleep health coefficient according to the user historical monitoring data, the user basic data, the historical medical data and the sleep data; a body health module for determining a body health coefficient according to the body movement data and the core body health data; an environment safety module for determining a medical environment safety coefficient according to the medical environment data; a monitoring report module for determining a monitoring report according to the sleep health coefficient, the body health coefficient and the medical environment safety coefficient.

[0006] According to the present application, the sleep health coefficient is determined according to the user historical monitoring data, the user basic data, the historical medical data and the sleep data, comprising: determining a mental health important result and a chronic disease management important result according to the historical medical data; if the user has a mental health disease, the mental health important result is 1, otherwise the mental health important result is 0; If suffering from a chronic disease, the chronic disease management important result is 1, otherwise, the chronic disease management important result is 0; According to the user basic data, the user age, the user medication amount and the user medication frequency are determined; According to the user historical monitoring data, the reference total sleep time and the reference sleep time are determined; According to the sleep data, the total sleep time, the sleep time, the wake-up times, the sleep efficiency and the deep sleep proportion are determined; The user age, the user medication amount, the user medication frequency, the reference total sleep time, the reference sleep time, the total sleep time, the sleep time, the wake-up times, the sleep efficiency and the deep sleep proportion are processed by the trained psychosomatic state prediction model to obtain a predicted psychosomatic state coefficient; The sleep health coefficient is determined by subtracting the mental health important result and the chronic disease management important result from the predicted psychosomatic state coefficient.

[0007] According to the present application, the training step of the psychosomatic state prediction model comprises: The historical other user age, the historical medication amount, the historical medication frequency, the historical total sleep time, the historical sleep time, the historical wake-up times, the historical sleep efficiency and the historical deep sleep proportion of other users in a plurality of historical test periods are obtained; According to the historical total sleep time and the historical sleep time, the historical reference total sleep time and the historical reference sleep time are determined; The historical reference total sleep time, the historical reference sleep time, the historical other user age, the historical medication amount, the historical medication frequency, the historical total sleep time, the historical sleep time, the historical wake-up times, the historical sleep efficiency and the historical deep sleep proportion are processed by the psychosomatic state prediction model to determine a sample predicted psychosomatic state coefficient; The questionnaire information of other users in a plurality of historical test periods is obtained; According to the questionnaire information, the historical actual psychosomatic state coefficient is determined; According to the sample predicted psychosomatic state coefficient, the historical actual psychosomatic state coefficient, the historical reference total sleep time, the historical reference sleep time, the historical other user age, the historical medication amount, the historical medication frequency, the historical total sleep time, the historical sleep time, the historical wake-up times, the historical sleep efficiency and the historical deep sleep proportion, the training loss function of the psychosomatic state prediction model is determined; The psychosomatic state prediction model is trained according to the training loss function of the psychosomatic state prediction model to obtain the trained psychosomatic state prediction model.

[0008] According to the present application, the historical actual psychosomatic state coefficient is determined according to the questionnaire information, comprising: According to the questionnaire information, the PHQ-9 questionnaire score, the survey working time and the survey exercise time are determined; According to the survey working time and the preset working time threshold, the first ratio is determined; According to the survey exercise time and the preset exercise time threshold, the second ratio is determined; According to the formula:

[0009] The historical actual psychosomatic state coefficient of the i-th other user in the j-th historical test cycle is determined , wherein, is the first ratio of the i-th other user in the j-th historical test cycle, is the second ratio of the i-th other user in the j-th historical test cycle, is the PHQ-9 questionnaire score of the i-th other user in the j-th historical test cycle.

[0010] According to the present application, the training loss function of the psychosomatic state prediction model is determined according to the sample predicted psychosomatic state coefficient, the historical actual psychosomatic state coefficient, the historical reference total sleep time, the historical reference sleep time, the historical user age, the historical drug amount, the historical drug frequency, the historical total sleep time, the historical sleep time, the historical wake-up times, the historical sleep efficiency and the historical deep sleep proportion, comprising: according to the formula:

[0011] The training loss function of the psychosomatic state prediction model is determined , wherein, is the sample predicted psychosomatic state coefficient of the i-th other user in the j-th historical test cycle, is the historical actual psychosomatic state coefficient of the i-th other user in the j-th historical test cycle, is the historical drug amount of the i-th other user in the j-th historical test cycle, is the preset drug amount threshold, is the historical drug frequency of the i-th other user in the j-th historical test cycle, is the preset drug frequency threshold, is the historical total sleep time of the i-th other user in the j-th historical test cycle, is the historical reference total sleep time of the i-th other user, a historical sleep time of the i th other user in a j th historical test period, a historical reference sleep time of the i th other user, a historical other user age of the i th other user in a j th historical test period, a preset age threshold, a historical wake-up times of the i th other user in a j th historical test period, a preset wake-up times threshold, a historical sleep efficiency of the i th other user in a j th historical test period, a preset sleep efficiency threshold, a historical deep sleep proportion of the i th other user in a j th historical test period, a preset historical deep sleep proportion threshold, n is a number of other users, i≤n, m is a number of historical test periods, j≤m, i, n, j and m are all positive integers.

[0012] According to the application, according to the body movement data and the core physical health data, a physical health coefficient is determined, comprising: determining a body movement condition according to the body movement data; determining a heart rate, a blood pressure, a blood oxygen saturation, a body temperature and a respiratory frequency according to the core physical health data; determining a set heart rate threshold, a set blood pressure threshold, a set blood oxygen saturation threshold, a set body temperature threshold and a set respiratory frequency threshold according to the body movement condition; judging whether the heart rate, the blood pressure, the blood oxygen saturation, the body temperature and the respiratory frequency conform to corresponding set safety thresholds according to the set heart rate threshold, the set blood pressure threshold, the set blood oxygen saturation threshold, the set body temperature threshold, the set respiratory frequency threshold, the heart rate, the blood pressure, the blood oxygen saturation, the body temperature and the respiratory frequency, if there is data not conforming to the corresponding set safety threshold, the physical health coefficient is 0, indicating that there is an abnormality, otherwise, the physical health coefficient is 1.

[0013] According to the application, according to the medical environment data, a medical environment safety coefficient is determined, comprising: determining air environment data according to the medical environment data, wherein the air environment data comprises a carbon dioxide concentration, a particulate matter concentration and a volatile organic compound concentration; determining an air environment safety coefficient according to the air environment data; If the carbon dioxide concentration is less than 800 ppm, the particulate matter concentration is less than 35 μg / m³, and the volatile organic matter concentration is less than 0.2 mg / m³, the air environment safety factor is 1, otherwise, the air environment safety factor is 0; According to the medical environment data, temperature data and humidity data are determined; According to the temperature data and the humidity data, a temperature and humidity environment safety factor is determined; If the temperature data is between 20 degrees Celsius and 25 degrees Celsius, and the humidity data is between 40% and 60%, the temperature and humidity environment safety factor is 2, otherwise, the temperature and humidity environment safety factor is 0; The air environment safety factor and the temperature and humidity environment safety factor are summed to determine a medical environment safety factor.

[0014] According to a second aspect of the present application, a patient closed-loop management method based on drug-treatment scheme risk stratification is provided, comprising: At multiple time points of a monitoring period, medical environment data, body movement data, sleep data, and core body health data are obtained through an intelligent medical terminal and sensors; User historical medical data, user basic data, and user historical monitoring data are obtained; According to the user historical monitoring data, the user basic data, the historical medical data, and the sleep data, a sleep health coefficient is determined; According to the body movement data and the core body health data, a body health coefficient is determined; According to the medical environment data, a medical environment safety factor is determined; According to the sleep health coefficient, the body health coefficient, and the medical environment safety factor, a monitoring report is determined.

[0015] Technical effects: According to the present application, medical environment data, user body movement data, sleep data and core body health data can be accurately collected, and sleep condition, physical condition and environmental condition are evaluated according to sleep data, core body health data and medical environment data respectively, sleep health coefficient, physical health coefficient and medical environment safety coefficient are determined, and monitoring report is generated, which improves the comprehensiveness and accuracy of patient closed-loop management. When determining the training loss function of the psychosomatic state prediction model, the psychosomatic state coefficient predicted by the sample, the historical actual psychosomatic state coefficient, the historical reference total sleep time, the historical reference sleep time, the historical drug amount, the historical drug frequency, the historical age of other users, the historical total sleep time, the historical sleep time, the historical wake-up times, the historical sleep efficiency and the historical deep sleep proportion are used to determine the training loss function of the psychosomatic state prediction model. In the calculation process, the possible influence of drug amount, drug frequency, age, total sleep time, sleep time, wake-up times, sleep efficiency and deep sleep proportion on psychosomatic state can be determined, and the influence of the above data on the error of the sample predicted psychosomatic state coefficient is determined based on the influence and the relative error of the sample predicted psychosomatic state coefficient. Set the training loss function, so that the psychosomatic state prediction model reduces the training loss function in the training process, and more accurately improves the accuracy of the psychosomatic state prediction model.

[0016] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, but not limiting the present application. Other features and aspects of the present application will become more apparent from the following detailed description of exemplary embodiments with reference to the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, the drawings needed in the embodiment or prior art description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can obtain other embodiments according to these drawings without creating any creative labor; Figure 1 Exemplarily show the block diagram of the patient closed-loop management system based on drug-treatment scheme risk stratification according to the embodiment of the present application; Figure 2 Exemplarily show the flowchart of the patient closed-loop management method based on drug-treatment scheme risk stratification according to the embodiment of the present application. DETAILED DESCRIPTION

[0018] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0019] The technical solutions of the present application will be described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and some embodiments may not be described again for the same or similar concepts or processes.

[0020] Figure 1 An example of a block diagram of a patient closed-loop management system based on drug-treatment plan risk stratification according to an embodiment of the present application is shown, which includes: A monitoring data module for acquiring medical environment data, body movement data, sleep data and core physical health data through an intelligent medical terminal and sensors at multiple time points of a monitoring period. A user data module for acquiring historical medical data of a user, basic data of the user and historical monitoring data of the user. A sleep health module for determining a sleep health coefficient according to the historical monitoring data of the user, the basic data of the user, the historical medical data and the sleep data. A physical health module for determining a physical health coefficient according to the body movement data and the core physical health data. An environment safety module for determining a medical environment safety coefficient according to the medical environment data. A monitoring report module for determining a monitoring report according to the sleep health coefficient, the physical health coefficient and the medical environment safety coefficient.

[0021] The patient closed-loop management system based on drug-treatment plan risk stratification according to the embodiments of the present application can accurately collect medical environment data and body movement data, sleep data and core physical health data of a user, and respectively evaluate sleep condition, physical condition and environment condition according to sleep data, core physical health data and medical environment data to determine sleep health coefficient, physical health coefficient and medical environment safety coefficient, generate a monitoring report, and improve the comprehensiveness and accuracy of patient closed-loop management.

[0022] According to an embodiment of the present application, in the monitoring data module, medical environment data, body movement data, sleep data and core physical health data are acquired through an intelligent medical terminal and sensors at multiple time points of a monitoring period.

[0023] For example, a monitoring period is set as 12h (e.g., 21:00-9:00), medical environment data (e.g., temperature and humidity data, carbon dioxide data) is acquired through gas sensors and temperature and humidity sensors arranged in the monitoring environment, body movement data (e.g., acceleration) and sleep data (e.g., total sleep time, deep sleep proportion) are acquired through a smart bracelet worn by the user, core physical health data (e.g., heart rate, blood pressure, blood oxygen saturation, body temperature and respiratory rate) is acquired through medical monitoring instruments and the smart bracelet, and the acquired data is stored in a database.

[0024] According to an embodiment of the present application, in the user data module, historical medical data of the user, basic data of the user and historical monitoring data of the user are acquired.

[0025] For example, the historical medical data (e.g., medical record) of the user is acquired through a medical system, the basic data (e.g., user age) of the user is acquired through a medical system, and the historical monitoring data of the user is acquired through a database.

[0026] According to an embodiment of the present application, in the sleep monitoring module, a sleep health coefficient is determined according to the historical monitoring data of the user, the basic data of the user, the historical medical data and the sleep data.

[0027] According to an embodiment of the present application, determining a sleep health coefficient according to the historical monitoring data of the user, the basic data of the user, the historical medical data and the sleep data comprises: determining a mental health important result and a chronic disease management important result according to the historical medical data; if the user has a mental health disease, the mental health important result is 1, otherwise, the mental health important result is 0; if the user has a chronic disease, the chronic disease management important result is 1, otherwise, the chronic disease management important result is 0; determining user age, user medication amount and user medication frequency according to the basic data of the user; determining a reference total sleep time and a reference sleep onset time according to the historical monitoring data of the user; determining a total sleep time, a sleep onset time, a wake-up frequency, a sleep efficiency and a deep sleep proportion according to the sleep data; processing the user age, the user medication amount, the user medication frequency, the reference total sleep time, the reference sleep onset time, the total sleep time, the sleep onset time, the wake-up frequency, the sleep efficiency and the deep sleep proportion through a trained physical and mental state prediction model to obtain a predicted physical and mental state coefficient; The sleep health coefficient is determined by subtracting the mental health important result and the chronic disease management important result from the predicted psychosomatic state coefficient.

[0028] For example, according to the medical record of the user, it is determined whether the user has a mental health disease and a chronic disease, when the user has a mental health disease and a chronic disease, the influence degree of the sleep condition on the recovery condition of the user is relatively greater, therefore, if the user has a mental health disease, the mental health important result is 1, otherwise, the mental health important result is 0, if the user has a chronic disease, the chronic disease management important result is 1, otherwise, the chronic disease management important result is 0; the user's age is acquired, the user's drug amount (the amount of a single drug use of the user) and the user's drug frequency in a monitoring period are acquired; according to the historical monitoring data of the user, the average total sleep time and the average sleep time of the user in a plurality of historical monitoring periods in the past are determined, that is, the reference total sleep time and the reference sleep time; according to the sleep data acquired by the smart bracelet, the total sleep time, the sleep time (the time required from turning off the light to prepare for sleep to actually falling asleep), the number of awakenings, the sleep efficiency (determined according to the ratio of the total sleep time to the total time in bed) and the proportion of deep sleep of the user in a monitoring period are determined; the user's age, the user's drug amount, the user's drug frequency, the reference total sleep time, the reference sleep time, the total sleep time, the sleep time, the number of awakenings, the sleep efficiency and the proportion of deep sleep are processed by the trained psychosomatic state prediction model to obtain the predicted psychosomatic state coefficient of the user on the next day, the psychosomatic state prediction model belongs to a kind of neural network model, and includes: data preprocessing and input layer, feature extraction layer, the psychosomatic state prediction model is trained by historical data, so that the psychosomatic state prediction model can generate a predicted psychosomatic state coefficient; the greater the predicted psychosomatic state coefficient, the better the psychosomatic condition of the next day after sleep, the healthier the sleep, the greater the mental health important result and the chronic disease management important result, indicating that the requirement degree of sleep health is higher, and the sleep health coefficient is determined by subtracting the mental health important result and the chronic disease management important result from the predicted psychosomatic state coefficient.

[0029] According to one embodiment of the present application, the training step of the psychosomatic state prediction model comprises: acquiring historical other user ages, historical drug amounts, historical drug frequencies, historical total sleep times, historical sleep times, historical number of awakenings, historical sleep efficiencies and historical proportions of deep sleep of other users in a plurality of historical test periods; determining historical reference total sleep times and historical reference sleep times according to the historical total sleep times and the historical sleep times; processing the historical reference total sleep time, the historical reference sleep time, the historical other user age, the historical drug amount, the historical drug frequency, the historical total sleep time, the historical sleep time, the historical wake-up times, the historical sleep efficiency and the historical deep sleep proportion through the psychosomatic state prediction model to determine a sample predicted psychosomatic state coefficient; obtaining survey questionnaire information of other users in a plurality of historical test cycles; determining a historical actual psychosomatic state coefficient according to the survey questionnaire information; determining a training loss function of the psychosomatic state prediction model according to the sample predicted psychosomatic state coefficient, the historical actual psychosomatic state coefficient, the historical reference total sleep time, the historical reference sleep time, the historical other user age, the historical drug amount, the historical drug frequency, the historical total sleep time, the historical sleep time, the historical wake-up times, the historical sleep efficiency and the historical deep sleep proportion; training the psychosomatic state prediction model according to the training loss function of the psychosomatic state prediction model to obtain a trained psychosomatic state prediction model.

[0030] For example, in a historical test cycle, the influence of sleep condition and drug condition on the next day's physical condition is tested, the historical drug amount, the historical drug frequency, the historical other user age, the historical total sleep time, the historical sleep time, the historical wake-up times, the historical sleep efficiency and the historical deep sleep proportion of other users in multiple historical test cycles are obtained, and the historical drug amount and the historical drug frequency set by medical personnel for other users in the medical order, that is, the preset drug amount threshold and the preset drug frequency threshold, are obtained; according to the average value of the historical total sleep time and the historical sleep time of other users in multiple historical test cycles, the historical reference total sleep time and the historical reference sleep time are determined; the historical drug amount, the historical drug frequency, the historical reference total sleep time, the historical reference sleep time, the historical other user age, the historical total sleep time, the historical sleep time, the historical wake-up times, the historical sleep efficiency and the historical deep sleep proportion are processed by the psychosomatic state prediction model to determine the sample predicted psychosomatic state coefficient; through the form of a questionnaire survey (such as a PHQ-9 questionnaire survey, and the questionnaire survey is only for the second day of the date corresponding to the historical test cycle), the psychological state of other users on the second day of the date corresponding to the historical test cycle is determined, and the questionnaire information of other users in multiple historical test cycles is obtained; according to the questionnaire information, the actual psychosomatic condition of other users on the second day of the date corresponding to the historical test cycle, that is, the historical actual psychosomatic state coefficient, is evaluated; according to the sample predicted psychosomatic state coefficient, the historical actual psychosomatic state coefficient, the historical drug amount, the historical drug frequency, the historical reference total sleep time, the historical reference sleep time, the historical other user age, the historical total sleep time, the historical sleep time, the historical wake-up times, the historical sleep efficiency and the historical deep sleep proportion, the training loss function of the psychosomatic state prediction model is determined; the psychosomatic state prediction model is trained according to the training loss function of the psychosomatic state prediction model, the accuracy of the prediction of the psychosomatic state prediction model is improved, and the trained psychosomatic state prediction model is obtained.

[0031] According to one embodiment of the application, according to the questionnaire information, the historical actual psychosomatic state coefficient is determined, comprising: According to the questionnaire information, the PHQ-9 questionnaire survey score, the survey working time and the survey exercise time are determined; According to the survey working time and the preset working time threshold, a first ratio is determined; According to the survey exercise time and the preset exercise time threshold, a second ratio is determined; According to formula (1), the historical actual psychosomatic state coefficient of the i-th other user in the j-th historical test cycle is determined , (1) wherein, a first ratio of the i-th other user in the j-th historical test period, a second ratio of the i-th other user in the j-th historical test period, a PHQ-9 questionnaire score of the i-th other user in the j-th historical test period.

[0032] For example, the PHQ-9 questionnaire is conducted on the other user to obtain a PHQ-9 questionnaire score, the PHQ-9 questionnaire score ranges from 0 to 27, and the greater the score, the worse the mental state. The questionnaire is conducted on the other user to determine the working time and the exercise time of the other user in the daytime after experiencing the test sleep, that is, the investigation working time and the investigation exercise time; the first ratio is determined according to the ratio of the investigation working time to a preset working time threshold, and the greater the first ratio, the more vigorous the energy of the other user after experiencing the sleep and the better the physical condition; the second ratio is determined according to the ratio of the investigation exercise time to a preset exercise time threshold, and the greater the second ratio, the more vigorous the energy of the other user after experiencing the sleep and the better the physical condition; the historical actual psychosomatic state coefficient is determined according to the PHQ-9 questionnaire score, the first ratio and the second ratio, for example, according to the formula: , the historical actual psychosomatic state coefficient of the i-th other user in the j-th historical test period is determined , wherein a first ratio of the i-th other user in the j-th historical test period, a second ratio of the i-th other user in the j-th historical test period, a PHQ-9 questionnaire score of the i-th other user in the j-th historical test period; the greater the first ratio, the more vigorous the energy of the other user after experiencing the sleep and the better the physical condition, the greater the second ratio, the more vigorous the energy of the other user after experiencing the sleep and the better the physical condition; the greater the PHQ-9 questionnaire score, the greater the first ratio, the greater the second ratio, and the better the mental state of the other user after experiencing the sleep, the greater the PHQ-9 questionnaire score.

[0033] According to one embodiment of the present application, the training loss function of the psychosomatic state prediction model is determined according to the sample predicted psychosomatic state coefficient, the historical actual psychosomatic state coefficient, the historical reference total sleep time, the historical reference sleep time, the historical other user age, the historical drug amount, the historical drug frequency, the historical total sleep time, the historical sleep time, the historical wake-up times, the historical sleep efficiency and the historical deep sleep proportion, including: the training loss function of the psychosomatic state prediction model is determined according to formula (2) , (2) wherein, is a sample predicted psychophysiological state coefficient of the i-th other user in the j-th historical test period, is a historical actual psychophysiological state coefficient of the i-th other user in the j-th historical test period, is a historical drug amount of the i-th other user in the j-th historical test period, is a preset drug amount threshold, is a historical drug frequency of the i-th other user in the j-th historical test period, is a preset drug frequency threshold, is a historical total sleep time of the i-th other user in the j-th historical test period, is a historical reference total sleep time of the i-th other user, is a historical sleep-in time of the i-th other user in the j-th historical test period, is a historical reference sleep-in time of the i-th other user, is a historical other user age of the i-th other user in the j-th historical test period, is a preset age threshold, is a historical wake-up times of the i-th other user in the j-th historical test period, is a preset wake-up times threshold, is a historical sleep efficiency of the i-th other user in the j-th historical test period, is a preset sleep efficiency threshold, is a historical deep sleep proportion of the i-th other user in the j-th historical test period, is a preset historical deep sleep proportion threshold, n is the number of other users, i≤n, m is the number of historical test periods, j≤m, i, n, j and m are all positive integers.

[0034] According to one embodiment of the present application, is a ratio of the historical sleep efficiency of the i-th other user in the j-th historical test period to the preset sleep efficiency threshold, the larger the ratio, the higher the historical sleep efficiency, may be set to 85%, is a ratio of the historical deep sleep proportion of the i-th other user in the j-th historical test period to the preset historical deep sleep proportion threshold, the larger the ratio, the larger the historical deep sleep proportion, may be set to 20%, a relative difference between the historical total sleep time of the i-th other user in the j-th historical test period and the historical reference total sleep time of the i-th other user, the larger the ratio, the longer or shorter the historical total sleep time of the i-th other user in the j-th historical test period, a relative difference between the historical medication amount of the i-th other user in the j-th historical test period and the preset medication amount threshold, the larger the ratio, the higher or lower the medication amount, the preset medication amount threshold being determined according to the medical order of the other user, a relative difference between the historical medication frequency of the i-th other user in the j-th historical test period and the preset medication frequency threshold, the larger the ratio, the higher or lower the medication frequency, the preset medication frequency threshold being determined according to the medical order of the other user, a ratio of the historical sleep time of the i-th other user in the j-th historical test period to the historical reference sleep time of the i-th other user, the larger the ratio, the longer the historical sleep time, a ratio of the historical other user age of the i-th other user in the j-th historical test period to the preset age threshold, the larger the ratio, the larger the historical other user age, which can be set to 30 years old, a ratio of the historical wake-up times of the i-th other user in the j-th historical test period to the preset wake-up times threshold, the larger the ratio, the more the historical wake-up times, which can be set to 1 time. the historical sleep efficiency, the historical deep sleep proportion and the sample predicted psychosomatic state coefficient are positively correlated, the historical drug amount, the relative difference between the historical drug frequency and the preset drug frequency threshold, the relative difference between the historical total sleep time and the historical baseline total sleep time, the historical sleep latency, the historical other user age and the historical wake-up times and the sample predicted psychosomatic state coefficient are negatively correlated, for example, when the historical sleep efficiency is higher, the sleep cycle (about 90-120 minutes per cycle) can be circularly completed 4-5 times, so that the body repair (deep sleep dominant) and the brain memory arrangement and emotion adjustment (REM sleep dominant) functions can be better exerted, the sample predicted psychosomatic state coefficient is larger, when the historical deep sleep proportion is higher, the physiological recovery effect is higher, the body condition after getting up is better, the sample predicted psychosomatic state coefficient is larger, when the relative difference between the historical drug amount and the preset drug amount threshold is larger, the drug of the other user is less standardized, the drug effect is poorer, the sample predicted psychosomatic state coefficient is smaller, when the relative difference between the historical drug frequency and the preset drug frequency threshold is larger, the drug of the other user is less standardized, the drug effect is poorer, the sample predicted psychosomatic state coefficient is smaller, when the relative difference between the historical total sleep time and the historical baseline total sleep time is larger, the historical total sleep time is relatively longer or shorter, the historical total sleep time is relatively shorter, which can cause the sleep structure to be compressed and damaged, that is, the body can preferentially ensure deep sleep (core repair function), but can sacrifice REM sleep (rapid eye movement sleep, related to emotion adjustment and memory integration), which means that although the core physiological function is maintained, the emotion and cognitive function are damaged, the sample predicted psychosomatic state coefficient is smaller, the historical total sleep time is relatively longer, which can cause sleep fragmentation and biological clock disorder, fatigue and sleepiness and cognitive impairment, the sample predicted psychosomatic state coefficient is smaller, when the historical sleep latency is longer, the sleep latency is longer, which can cause sleep fragmentation, fatigue and sleepiness and emotional irritability and irritability, the sample predicted psychosomatic state coefficient is smaller, when the historical other user age is older, the sleep quality is generally lower, the sample predicted psychosomatic state coefficient is smaller, when the historical wake-up times are more, the sleep cycle is interrupted more times, which can cause daytime fatigue and sleepiness and emotional disorder, the sample predicted psychosomatic state coefficient is smaller, therefore, the historical sleep efficiency and the historical deep sleep proportion related terms are placed in the numerator, the historical drug amount, the historical drug frequency, the historical total sleep time, the historical sleep latency, the historical other user age and the historical wake-up times related terms are placed in the denominator, the values of and are larger, the value of the sample predicted psychosomatic state coefficient is larger, the error influence on the sample predicted psychosomatic state coefficient is larger, , , , , and The smaller the value, the larger the value of the sample predicted mental and physical state coefficient, and the greater the impact on the error of the sample predicted mental and physical state coefficient.

[0035] According to one embodiment of the present invention, To predict the relative error between the physical and mental state coefficient of the i-th other user in the j-th historical test period and the actual historical physical and mental state coefficient, we use... The training loss function is determined by weighted averaging the relative errors between the predicted mental and physical state coefficients of the i-th other user in the j-th historical test period and the actual historical mental and physical state coefficients. During training, the training loss function is reduced to decrease the error between the predicted mental and physical state coefficients and the actual historical mental and physical state coefficients, thereby improving the accuracy of the mental and physical state prediction model.

[0036] In this way, the training loss function of the physical and mental state prediction model can be determined based on the sample predicted physical and mental state coefficient, historical actual physical and mental state coefficient, historical baseline total sleep time, historical baseline sleep onset time, historical medication dosage, historical medication frequency, historical age of other users, historical total sleep time, historical sleep onset time, historical number of awakenings, historical sleep efficiency, and historical deep sleep percentage. During the calculation process, the possible impact of medication dosage, medication frequency, age, total sleep time, sleep onset time, number of awakenings, sleep efficiency, and deep sleep percentage on physical and mental state can be used to determine the impact of the above data on the error of the sample predicted physical and mental state coefficient. Based on this impact and the relative error of the sample predicted physical and mental state coefficient, the training loss function can be set to reduce the training loss function during the training process of the physical and mental state prediction model, thereby improving the accuracy of the physical and mental state prediction model in a more targeted manner.

[0037] According to one embodiment of the present invention, in the physical health module, a physical health coefficient is determined based on the body movement data and the core physical health data.

[0038] According to one embodiment of the present invention, determining the physical health coefficient based on the body movement data and the core physical health data includes: Based on the aforementioned body movement data, determine the body movement status; Based on the aforementioned core health data, determine heart rate, blood pressure, blood oxygen saturation, body temperature, and respiratory rate; Based on the described physical activity status, determine and set the heart rate threshold, blood pressure threshold, blood oxygen saturation threshold, body temperature threshold, and respiratory rate threshold. According to the set heart rate threshold, the set blood pressure threshold, the set blood oxygen saturation threshold, the set body temperature threshold, the set respiratory rate threshold, the heart rate, the blood pressure, the blood oxygen saturation, the body temperature and the respiratory rate, it is judged whether the heart rate, the blood pressure, the blood oxygen saturation, the body temperature and the respiratory rate conform to the corresponding set safety threshold, if there is data not conforming to the corresponding set safety threshold, the physical health coefficient is 0, indicating that there is an abnormality, otherwise, the physical health coefficient is 1.

[0039] For example, according to the body motion data collected by the smart bracelet, the body motion condition of the user is determined, such as sleep, stillness (awake, resting state) and motion; the heart rate, blood pressure, blood oxygen saturation, body temperature and respiratory rate of the user are collected by the smart bracelet; according to the body motion condition, the set heart rate threshold, the set blood pressure threshold, the set blood oxygen saturation threshold, the set body temperature threshold and the set respiratory rate threshold are determined, such as the threshold values corresponding to the heart rate, blood pressure, blood oxygen saturation, body temperature and respiratory rate in the sleep condition are 40-60 times / minute, 90-115 mmHg / 60-75 mmHg, greater than or equal to 95%, 36-36.5 degrees Celsius, 12-16 times / minute, the threshold values corresponding to the heart rate, blood pressure, blood oxygen saturation, body temperature and respiratory rate in the stillness condition are 60-100 times / minute, 120-139 mmHg / 80-89 mmHg, greater than or equal to 95%, 36.3-37.2 degrees Celsius, 12-20 times / minute, and the threshold values corresponding to the heart rate, blood pressure, blood oxygen saturation, body temperature and respiratory rate in the motion condition are less than or equal to (220 minus age)*0.85 times / minute, 160-220 mmHg / 80-100 mmHg, greater than or equal to 95%, 36.3-38.5 degrees Celsius, 40-50 times / minute; according to the set heart rate threshold, the set blood pressure threshold, the set blood oxygen saturation threshold, the set body temperature threshold, the set respiratory rate threshold, the heart rate, the blood pressure, the blood oxygen saturation, the body temperature and the respiratory rate, it is judged whether the heart rate, the blood pressure, the blood oxygen saturation, the body temperature and the respiratory rate conform to the corresponding set safety threshold, if there is data not conforming to the corresponding set safety threshold, the physical health coefficient is 0, indicating that there is an abnormality, otherwise, the physical health coefficient is 1.

[0040] According to one embodiment of the present application, in the environmental safety module, according to the medical environment data, a medical environment safety coefficient is determined.

[0041] According to one embodiment of the present application, determining a medical environment safety coefficient according to the medical environment data comprises: determining air environment data according to the medical environment data, wherein the air environment data comprises carbon dioxide concentration, particulate matter concentration and volatile organic compound concentration; determining air environment safety factor according to the air environment data; if the carbon dioxide concentration is less than 800ppm, the particulate matter concentration is less than 35μg / m³ and the volatile organic compound concentration is less than 0.2mg / m³, the air environment safety factor is 1, otherwise, the air environment safety factor is 0; determining temperature data and humidity data according to the medical environment data; determining temperature and humidity environment safety factor according to the temperature data and the humidity data; if the temperature data is between 20-25℃ and the humidity data is between 40%-60%, the temperature and humidity environment safety factor is 2, otherwise, the temperature and humidity environment safety factor is 0; summing the air environment safety factor and the temperature and humidity environment safety factor to determine the medical environment safety factor.

[0042] For example, the carbon dioxide concentration, the particulate matter concentration (PM2.5 concentration) and the volatile organic compound concentration (formaldehyde concentration) in the air are detected by a gas sensor; if the carbon dioxide concentration is less than 800ppm, the particulate matter concentration is less than 35μg / m³ and the volatile organic compound concentration is less than 0.2mg / m³, the air environment safety factor is 1, otherwise, the air environment safety factor is 0; the temperature data is detected by a temperature and humidity sensor; if the temperature data is between 20-25℃ (to avoid cardiovascular stress caused by overcooling or overheating) and the humidity data is between 40%-60% (to avoid dryness leading to respiratory tract infection, humidity breeding mold and mites and causing allergy), the temperature and humidity environment safety factor is 2, otherwise, the temperature and humidity environment safety factor is 0; the medical environment safety factor is determined by summing the air environment safety factor and the temperature and humidity environment safety factor.

[0043] According to one embodiment of the present application, in the monitoring report module, a monitoring report is determined according to the sleep health factor, the physical health factor and the medical environment safety factor.

[0044] For example, if the sleep health coefficient is greater than or equal to 16, it indicates that the user's sleep condition is good, if the sleep health coefficient is less than 16, it indicates that the user's sleep condition is poor, and it may have a negative impact on the subsequent mood and physical condition, the user is reminded when the user gets up, if the physical health coefficient is 0, it indicates that the user's heart rate or blood pressure or blood oxygen saturation or body temperature or respiratory rate is abnormal, the relevant staff is notified, if the physical health coefficient is 1, the monitoring is continued, if the medical environment safety coefficient is 3, it indicates that the medical environment is in normal condition, the monitoring is continued, if the medical environment safety coefficient is 2, it indicates that the air exists abnormal condition, the user is reminded to take ventilation measures, if the medical environment safety coefficient is 1, it indicates that the temperature and humidity exist abnormal condition, the air conditioning device (such as air conditioner) is started to adjust, if the medical environment safety coefficient is 0, it indicates that the air and the temperature and humidity exist abnormal condition, the user and the staff are notified, and corresponding measures are taken.

[0045] The patient closed-loop management system based on drug-treatment scheme risk stratification according to the embodiment of the application can accurately collect medical environment data and user's body movement data, sleep data and core physical health data, and respectively evaluate sleep condition, physical condition and environment condition according to the sleep data, the core physical health data and the medical environment data, determine sleep health coefficient, physical health coefficient and medical environment safety coefficient, and generate monitoring report, which improves the comprehensiveness and accuracy of patient closed-loop management. When determining the training loss function of the mind-body state prediction model, the training loss function of the mind-body state prediction model can be determined according to the sample predicted mind-body state coefficient, the historical actual mind-body state coefficient, the historical baseline total sleep time, the historical baseline sleep time, the historical drug amount, the historical drug frequency, the historical age of other users, the historical total sleep time, the historical sleep time, the historical wake-up times, the historical sleep efficiency and the historical deep sleep proportion. In the calculation process, the influence of the drug amount, the drug frequency, the age, the total sleep time, the sleep time, the wake-up times, the sleep efficiency and the deep sleep proportion on the mind-body state can be determined, and the influence of the above data on the error of the sample predicted mind-body state coefficient is determined, and the training loss function is set based on the influence and the relative error of the sample predicted mind-body state coefficient, so that the mind-body state prediction model reduces the training loss function in the training process, and more accurately improves the accuracy of the mind-body state prediction model.

[0046] Figure 2 An exemplary flowchart of a patient closed-loop management method based on drug-treatment scheme risk stratification according to an embodiment of the application is shown, which comprises: In step S1, at multiple time points in a monitoring period, medical environment data, body movement data, sleep data and core physical health data are acquired through an intelligent medical terminal and sensors. Step S2, obtaining historical medical data of the user, basic data of the user and historical monitoring data of the user; Step S3, determining a sleep health coefficient according to the historical monitoring data of the user, the basic data of the user, the historical medical data and the sleep data; Step S4, determining a physical health coefficient according to the body movement data and the core physical health data; Step S5, determining a medical environment safety coefficient according to the medical environment data; Step S6, determining a monitoring report according to the sleep health coefficient, the physical health coefficient and the medical environment safety coefficient.

[0047] The present application can be a method, apparatus, system and / or computer program product. The computer program product can include a computer-readable storage medium having computer-readable program instructions loaded thereon for executing various aspects of the present application.

[0048] Those skilled in the art will understand that the above-described embodiments of the present application shown in the description and drawings are only as examples and do not limit the present application. The purpose of the present application has been fully and effectively achieved. The functional and structural principles of the present application have been demonstrated and described in the embodiments, and the embodiments of the present application can be modified or changed in any way without departing from the principles.

Claims

1. A patient closed loop management system based on drug-therapy regimen risk stratification, characterized in that, The method comprises the following steps: monitoring data module, for acquiring medical environment data, body movement data, sleep data and core physical health data through the intelligent medical terminal and the sensor at multiple time points of the monitoring period; user data module, for acquiring historical medical data, basic data and historical monitoring data of the user; sleep health module, for determining a sleep health coefficient according to the historical monitoring data, the basic data, the historical medical data and the sleep data of the user; physical health module, for determining a physical health coefficient according to the body movement data and the core physical health data; environment safety module, for determining a medical environment safety coefficient according to the medical environment data; monitoring report module, for determining a monitoring report according to the sleep health coefficient, the physical health coefficient and the medical environment safety coefficient.

2. The patient closed loop management system based on drug-treatment regimen risk stratification of claim 1, wherein, determining a sleep health coefficient according to the historical monitoring data, the basic data, the historical medical data and the sleep data of the user comprises the following steps: determining a mental health important result and a chronic disease management important result according to the historical medical data; if the user has a mental health disease, the mental health important result is 1, otherwise, the mental health important result is 0; if the user has a chronic disease, the chronic disease management important result is 1, otherwise, the chronic disease management important result is 0; determining the user's age, the user's medication amount and the user's medication frequency according to the basic data of the user; determining a baseline total sleep time and a baseline sleep time according to the historical monitoring data of the user; determining a total sleep time, a sleep time, a wake-up frequency, a sleep efficiency and a deep sleep proportion according to the sleep data; processing the user's age, the user's medication amount, the user's medication frequency, the baseline total sleep time, the baseline sleep time, the total sleep time, the sleep time, the wake-up frequency, the sleep efficiency and the deep sleep proportion through a trained physical and mental state prediction model to obtain a predicted physical and mental state coefficient; determining a sleep health coefficient by subtracting the mental health important result and the chronic disease management important result from the predicted physical and mental state coefficient.

3. The patient closed loop management system based on drug-treatment regimen risk stratification of claim 2, wherein, The training steps of the physical and mental state prediction model comprise the following steps: acquiring historical other user's age, historical medication amount, historical medication frequency, historical total sleep time, historical sleep time, historical wake-up frequency, historical sleep efficiency and historical deep sleep proportion of other users in multiple historical test periods; determining a historical baseline total sleep time and a historical baseline sleep time according to the historical total sleep time and the historical sleep time; processing the historical baseline total sleep time, the historical baseline sleep time, the historical other user's age, the historical medication amount, the historical medication frequency, the historical total sleep time, the historical sleep time, the historical wake-up frequency, the historical sleep efficiency and the historical deep sleep proportion through the physical and mental state prediction model to determine a sample predicted physical and mental state coefficient; acquiring questionnaire information of other users in multiple historical test periods; determining a historical actual physical and mental state coefficient according to the questionnaire information. According to the sample predicted psychosomatic state coefficient, the historical actual psychosomatic state coefficient, the historical reference total sleep time, the historical reference sleep time, the historical other user age, the historical drug amount, the historical drug frequency, the historical total sleep time, the historical sleep time, the historical wake-up times, the historical sleep efficiency and the historical deep sleep proportion, a training loss function of a psychosomatic state prediction model is determined. According to the training loss function of the psychosomatic state prediction model, the psychosomatic state prediction model is trained to obtain a trained psychosomatic state prediction model.

4. The patient closed loop management system based on drug-treatment regimen risk stratification of claim 3, wherein, According to the survey questionnaire information, a historical actual psychosomatic state coefficient is determined, including: According to the survey questionnaire information, a PHQ-9 questionnaire survey score, a survey working time and a survey exercise time are determined. According to the survey working time and a preset working time threshold, a first ratio is determined. According to the survey exercise time and a preset exercise time threshold, a second ratio is determined. According to the formula: determining a historical actual psychophysical state coefficient of the ith other user at a jth historical test period wherein, is a first ratio of the ith other user at the jth historical test period, is a second ratio of the ith other user at the jth historical test period, is a PHQ-9 questionnaire score of the ith other user at the jth historical test period.

5. The patient closed loop management system based on drug-treatment regimen risk stratification of claim 3, wherein, According to the sample predicted psychosomatic state coefficient, the historical actual psychosomatic state coefficient, the historical reference total sleep time, the historical reference sleep time, the historical other user age, the historical drug amount, the historical drug frequency, the historical total sleep time, the historical sleep time, the historical wake-up times, the historical sleep efficiency and the historical deep sleep proportion, a training loss function of a psychosomatic state prediction model is determined, including: according to the formula: Training loss function of psychosomatic state prediction model wherein, is a sample predicted psychosomatic state coefficient of the i th other user in the j th historical test period, is a historical actual psychosomatic state coefficient of the i th other user in the j th historical test period, is a historical drug amount of the i th other user in the j th historical test period, is a preset drug amount threshold, is a historical drug frequency of the i th other user in the j th historical test period, is a preset drug frequency threshold, is a historical total sleep time of the i th other user in the j th historical test period, is a historical reference total sleep time of the i th other user, is a historical sleep time of the i th other user in the j th historical test period, is a historical reference sleep time of the i th other user, is a historical other user age of the i th other user in the j th historical test period, is a preset age threshold, is a historical wake-up times of the i th other user in the j th historical test period, is a preset wake-up times threshold, is a historical sleep efficiency of the i th other user in the j th historical test period, is a preset sleep efficiency threshold, is a historical deep sleep proportion of the i th other user in the j th historical test period, is a preset historical deep sleep proportion threshold, n is the number of other users, i≤n, m is the number of historical test periods, j≤m, i, n, j and m are all positive integers.

6. The patient closed loop management system based on drug-treatment regimen risk stratification of claim 1, wherein, According to the body movement data and the core physical health data, a physical health coefficient is determined, including: According to the body movement data, a body movement condition is determined. According to the core physical health data, a heart rate, a blood pressure, a blood oxygen saturation, a body temperature and a respiratory frequency are determined. According to the body movement condition, a set heart rate threshold, a set blood pressure threshold, a set blood oxygen saturation threshold, a set body temperature threshold and a set respiratory frequency threshold are determined. According to the set heart rate threshold, the set blood pressure threshold, the set blood oxygen saturation threshold, the set body temperature threshold, the set respiratory frequency threshold, the heart rate, the blood pressure, the blood oxygen saturation, the body temperature and the respiratory frequency, it is judged whether the heart rate, the blood pressure, the blood oxygen saturation, the body temperature and the respiratory frequency meet the corresponding set safety threshold, if there is data that does not meet the corresponding set safety threshold, the physical health coefficient is 0, indicating that there is an anomaly, otherwise, the physical health coefficient is 1.

7. The patient closed loop management system based on drug-treatment regimen risk stratification of claim 1, wherein, According to the medical environment data, a medical environment safety coefficient is determined, including: According to the medical environment data, air environment data is determined, wherein the air environment data includes: carbon dioxide concentration, particulate matter concentration and volatile organic compound concentration; According to the air environment data, an air environment safety coefficient is determined. If the carbon dioxide concentration is less than 800 ppm, the particulate matter concentration is less than 35 μg / m³ and the volatile organic compound concentration is less than 0.2 mg / m³, the air environment safety coefficient is 1, otherwise, the air environment safety coefficient is 0; According to the medical environment data, temperature data and humidity data are determined; According to the temperature data and the humidity data, determine a temperature and humidity environment safety coefficient; If the temperature data is between 20 degrees Celsius and 25 degrees Celsius, and the humidity data is between 40% and 60%, the temperature and humidity environment safety coefficient is 2, otherwise, the temperature and humidity environment safety coefficient is 0; Sum the air environment safety coefficient and the temperature and humidity environment safety coefficient to determine a medical environment safety coefficient.

8. A patient closed-loop management method based on drug-treatment regimen risk stratification, the method is used for the patient closed-loop management system based on drug-treatment regimen risk stratification in any one of claims 1-7, characterized in that, Comprise: At multiple time points in a monitoring period, obtain medical environment data, body movement data, sleep data, and core body health data through an intelligent medical terminal and sensors; Obtain historical medical data of a user, basic data of the user, and historical monitoring data of the user; According to the user historical monitoring data, the user basic data, the historical medical data, and the sleep data, determine a sleep health coefficient; According to the body movement data and the core body health data, determine a body health coefficient; According to the medical environment data, determine a medical environment safety coefficient; According to the sleep health coefficient, the body health coefficient, and the medical environment safety coefficient, determine a monitoring report.

Images