Real-time management system for precise medication of co-patient with chronic disease based on AI (artificial intelligence) large model

Abstract

The invention provides a chronic disease co-patient accurate medication real-time management system based on an AI large model, and the system comprises accurate medication guided by the AI large model, through the mathematical modeling and simulation technology, the related information of patients, medicines and diseases is integrated, and accurate medication is guided. According to the method, the administration scheme can be formulated more accurately, the curative effect is improved, side effects are reduced, a patient medication report is provided for physicians and pharmacists, the physicians and pharmacists can dynamically master patient medication information in real time, reference is provided for the physicians to formulate the next medication scheme, and the treatment effect on chronic disease patients is improved.

Description

Technical Field

[0001] The present invention relates to the field of AI big model application and medical device technology, and in particular to a real-time management system for precise medication of patients with chronic comorbidities based on AI big model. Background Art

[0002] Chronic diseases refer to diseases that persist for a long time or develop over time. Common chronic diseases include cardiovascular and cerebrovascular diseases (such as hypertension, coronary heart disease, stroke, etc.), diabetes, malignant tumors, chronic obstructive pulmonary disease, and neurodegenerative diseases. The prevalence of chronic disease comorbidity among Chinese residents is obvious, with a prevalence of 36.3%, and the prevalence increases with age, especially in people between 60 and 80 years old. Chronic diseases are characterized by high disability rates and require long-term treatment and care, which imposes a heavy economic burden on patients and their families.

[0003] As people age, the number of patients with chronic diseases increases. A patient may have hypertension, diabetes, hyperlipidemia, hyperuricemia, heart failure and other basic diseases at the same time. They need to take a wide variety of drugs. When can each drug be taken to achieve the maximum efficacy? What are the dietary taboos during medication? How should lifestyle be adjusted? There are interactions between drugs, which will enhance or weaken the efficacy of another drug. How should patients adjust the dosage of drugs when taking them? How to minimize adverse reactions? If a drug with a narrow safe dosage is taken, how to monitor the blood concentration of the drug in real time. These are many problems that bother patients and doctors in life and clinical practice. In addition, patients who have just been diagnosed or found to have chronic diseases (such as hypertension) have low medication compliance and are worried that they will need to take medication for life once they start taking medication. They do not fully understand the hazards of hypertension, resulting in poor treatment compliance; doctors and pharmacists lack the channels and energy to carry out large-scale patient education, and it is difficult to remind patients to take medication on time. Some hypertensive patients go home or forget to take medication after the hospital prescribes medication, or take medication intermittently (not taking medication when blood pressure is normal, but taking medication only when blood pressure is high), which eventually leads to blood pressure fluctuations and poor control. When they go to the hospital for follow-up, some patients will lie when the doctor asks them about their condition, saying that they take medication on time when they do not, which misleads the doctor and is not conducive to the doctor prescribing accurate drug treatment plans. Personalized medication for patients with chronic comorbidities is a complex and multidimensional issue, which requires comprehensive consideration of factors such as the patient's specific situation, drug interactions, and individual differences. Elderly patients with chronic comorbidities often need to take multiple medications at the same time, which increases the risk of drug interactions and may lead to adverse reactions (ADRs). Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides a real-time management system for accurate medication of patients with chronic diseases and comorbidities based on an AI big model, which includes: a smart medicine box, a patient smart terminal APP, a physician smart terminal APP, a pharmacist smart terminal APP and a cloud data processing center;

[0005] The smart medicine box includes a medicine box and a smart sensor arranged in the medicine box. The medicine box is used to store medicines for treating different chronic diseases. The smart sensor is used to sense the changes in the environmental parameters in the box and the quantity and type of medicines in the box in real time, and generate a medication details report, and send the medication details report to the patient smart terminal APP and the cloud data processing center;

[0006] The patient smart terminal APP is used to set the medication time for different medicines; it is also used to receive the medication details report sent by the smart sensor; it is also used to obtain the patient's basic information, physical examination data, past medical history and all the medicines currently being taken, and upload the above information to the cloud data processing center;

[0007] The cloud data processing center is used to analyze potential drug interaction risks, medication precautions of various drugs and give diet and lifestyle recommendations based on the patient's basic information, physical examination data, past medical history and all medication information currently being taken, using an AI large model, and ultimately generates safe medication guidance information and sends it to the patient's smart terminal APP in real time to guide the patient's safe medication use. The safe medication guidance information includes real-time medication reminders, drug interaction risk reminders, medication precautions, dietary taboos and lifestyle guidance; it is also used to forward the patient's medication details report uploaded by the smart sensor to the physician's smart terminal APP and the pharmacist's smart terminal APP in real time.

[0008] Furthermore, the AI ​​big model adopts a twin graph neural network algorithm based on path perception, and generates edge feature tensors of the shortest reaction path distances between multiple path-aware drugs through multiple relay path sampling, thereby fusing the twin graph neural network to predict drug interactions.

[0009] Furthermore, the method for constructing the AI ​​big model includes:

[0010] Step S1: calling the data in the drug interaction database, collecting the drug instructions of commonly used drugs for chronic diseases and the medication data, physical examination data and medical history of several patients with chronic diseases;

[0011] Step S2: Select the Transformer neural network structure and use the deep learning framework TensorFlow to build a neural network model. The model structure is: a large language model LLM composed of approximately 70k multi-head self-attention modules as the main modules, with a total of 50 billion parameters; use cross-validation and grid search methods to optimize the model parameters. During the training process, the network weights are updated through the back-propagation algorithm, and the model performance is evaluated using the validation set, and the model parameters are continuously adjusted to achieve the optimal state.

[0012] Furthermore, the real-time management system for precise medication of patients with chronic diseases and comorbidities based on the AI ​​big model further includes a wearable module. The wearable module adopts a microneedle E-AB sensor array, inserts tiny needles through the surface of the skin, extracts samples directly from sweat or blood, and monitors the blood concentration of low-safety dose drugs in real time. The blood concentration of low-safety dose drugs is transmitted to the patient's smart terminal APP via Bluetooth for real-time display, and is uploaded to the cloud data processing center through the smart terminal APP, and then forwarded by the cloud data processing center to the physician smart terminal APP and the pharmacist smart terminal APP.

[0013] Furthermore, the intelligent sensor includes a humidity sensor, a temperature sensor, a position balance sensor, a micro-weight sensor, a light intensity sensor, a central processing module, a timer, a parameter setting module, an alarm module, a power module, a wireless communication module and a switch controller; the switch controller is used to record the time after the medicine box is opened and closed, and notify other sensors to record the environmental parameters and drug information of the medicine at that time; the central processing module is used to automatically identify the type of medicine that has been taken, the number of medicines and the medication time of each medicine, as well as the type and quantity of remaining medicines according to the weight changes of all medicines, and generate a detailed report on the patient's medication and send it to the cloud data processing center and the patient's smart terminal APP through the wireless communication module. The central processing module is also used to control the alarm module to remind the medication time in real time, and the central processing module is also used to determine whether the temperature, humidity and position balance exceed the threshold; the parameter setting module is used to set the threshold of the appropriate storage environment parameters of each medicine and the judgment threshold of the box being in a position balance state; the alarm module is used to automatically issue an alarm message when the temperature, humidity and position balance exceed the threshold or when the medication time of any medicine is reached.

[0014] Furthermore, the micro-weight sensor adopts an NFC smart sensor chip, which is packaged under each medicine position on the medicine plate. The sensor in the chip detects the state change in real time and uploads it to the central processing module for processing.

[0015] Furthermore, the medicine box includes a box body, an inner frame and a cover plate, the cover plate is connected to the upper part of the box body by a hinge, the inner lower part of the box body is provided with a bottom plate, the bottom plate is provided with a tablet medicine plate and a liquid medicine placement plate, the tablet medicine plate is provided with a micro-weight sensor, the liquid medicine placement plate is embedded with a position balance sensor, the cover plate is embedded with a detection device composed of a temperature sensor, a humidity sensor and a light intensity sensor, the detection device is located at the upper part of the tablet medicine plate and the liquid medicine placement plate, a partition plate is provided between the tablet medicine plate and the liquid medicine placement plate, the inner frame is provided with connecting parts on both sides and the lower part of the partition plate, the connecting parts A limiting shaft is provided on the connecting part, and limiting parts are provided on both sides and the lower part of the partition plate. The limiting shaft passes through the interior of the limiting part. A mounting part is also provided on the inner side of the inner frame at the lower part of the partition plate, and a screw is rotatably installed on the inner side of the mounting part. The screw is threadedly connected to the middle part of the lower side of the partition plate, one end of the screw is connected to a driving motor, and the output end of the driving motor is connected to one end of the screw through a coupling, and the driving motor is connected to the inner frame. A movable groove is opened at the lower part of the bottom plate, and the lower part of the partition plate and the limiting part located at the lower part of the partition plate are located inside the movable groove, and the width of the movable groove is matched with the width of the limiting part.

[0016] Furthermore, the intelligent terminal includes a mobile terminal and a PC terminal.

[0017] Compared with the prior art, the present invention has the following beneficial effects:

[0018] The real-time management system for accurate medication of chronic disease comorbidity patients based on AI big model provided by the present invention integrates relevant information of patients, drugs and diseases through big model, artificial intelligence neural network algorithm and deep learning and other technologies, analyzes potential drug interaction risks and medication precautions of various drugs through AI big model, gives diet and lifestyle suggestions, guides patients to use drugs safely, tailors the best personalized medication plan for chronic disease comorbidity patients, and guides chronic disease comorbidity patients to use drugs accurately and safely in real time, improves efficacy and reduces side effects. In addition, through intelligent sensors (such as temperature and humidity sensors, position balance sensors, chips, etc.), data such as drug storage environment and patient medication status are sensed, and patient medication status reports are provided to doctors and pharmacists, so that doctors and pharmacists can dynamically control the actual medication information of patients in real time, provide reference for doctors to formulate the next medication plan, and facilitate doctors to formulate more accurate medication plans for patients, and improve the treatment effect of chronic disease comorbidity patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a structural diagram of the real-time management system for precise medication of patients with chronic comorbidities based on the AI ​​big model of this application.

[0020] Figure 2 is a schematic diagram of an LLM neuron.

[0021] Figure 3 The present invention provides an overall external schematic diagram of an AI big model based on an AI big model for a real-time precise medication management system for patients with chronic diseases and comorbidities.

[0022] Figure 4 It is a schematic diagram of the internal assembly of an AI big model based on the AI ​​big model of the present invention, which is a real-time management system for precise medication of patients with chronic diseases and comorbidities.

[0023] Figure 5 It is a schematic top view of the interior of an AI big model of the present invention, which is based on the AI ​​big model and a real-time management system for precise medication of patients with chronic diseases and comorbidities.

[0024] Figure 6 It is a schematic diagram of the internal framework of an AI big model based on the AI ​​big model of the present invention for a real-time management system for precise medication of patients with chronic diseases and comorbidities.

[0025] Figure 7 This is a schematic diagram of the assembly of the cover plate part of the real-time management system for precise medication of patients with chronic diseases and comorbidities based on the AI ​​big model of the present invention.

[0026] In the figure, the corresponding relationship between the component names and the figure numbers is as follows:

[0027] 1. Box body; 2. Inner frame; 3. Cover plate; 4. Tablet plate; 5. Liquid medicine placement plate; 6. Liquid medicine bottle placement area; 7. Position balance sensor; 8. Tablet placement area; 9. Micro-weight sensor; 10. Partition plate; 11. Connector; 12. Limit shaft; 13. Limit member; 14. Mounting part; 15. Screw; 16. Drive motor; 17. Bottom plate; 18. Temperature sensor; 19. Humidity sensor; 20. Light intensity sensor; 21. Detection head; 22. Temperature display device; 23. Humidity display device; 24. Light intensity display device; 25. Alarm device; 26. Detection device; 27. Moving slot. DETAILED DESCRIPTION

[0028] The following is a further detailed description of the embodiments of the present invention in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention but are not intended to limit the scope of the present invention.

[0029] In the description of the present invention, unless otherwise specified, "plurality" means two or more than two; the orientations or positional relationships indicated by the terms "upper", "lower", "left", "right", "inner", "outer", "front end", "rear end", "head", "tail", etc. are based on the orientations or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", "third", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0030] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "connected" and "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0031] like Figure 1 As shown, this embodiment provides a real-time management system for precise medication of patients with chronic diseases and comorbidities based on an AI big model, which includes: a smart medicine box, a patient smart terminal APP, a physician smart terminal APP, a pharmacist smart terminal APP and a cloud data processing center.

[0032] The smart terminal includes a mobile terminal and a PC terminal. For example, the patient smart terminal APP includes a mobile terminal APP, a tablet APP, and a PC terminal APP, etc., the physician smart terminal APP includes a mobile terminal APP, a tablet APP, and a PC terminal APP, etc., and the pharmacist smart terminal APP includes a mobile terminal APP, a tablet APP, and a PC terminal APP, etc. The physician APP, the pharmacist APP, the physician PC client, the pharmacist PC client, and the patient APP share data information through the cloud data processing center to achieve accurate tracking and management of the medication of hypertensive patients by physicians and pharmacists.

[0033] The patient's smart terminal APP is used to set the medication time for different medicines; it is also used to receive the medication details report sent by the micro-weight sensor. It is also used to obtain the patient's basic information (such as name, height, weight, age, etc.), physical examination data (such as blood sugar, blood lipids, blood pressure, liver and kidney function, blood routine, urine routine, electrolytes, etc.), past medical history and all medication information currently being taken, and upload the above information to the cloud data processing center. The patient APP sets a medication reminder, and the instruction is sent to the smart sensor 9 to remind the patient to take the medication on time and improve the patient's medication compliance.

[0034] The intelligent medicine box includes a medicine box 1 and an intelligent sensor 9 arranged in the medicine box 1. The intelligent sensor 9 includes a humidity sensor 19, a temperature sensor 18, a position balance sensor 7, a micro-weight sensor, a light intensity sensor, a central processing module, a timer, a parameter setting module, an alarm module, a power module, a wireless communication module and a switch controller. The switch controller is used to record the time after the medicine box is opened and closed, and notify other sensors to record the environmental parameters and medicine information of the medicine at that time; the central processing module is used to automatically identify the type of medicine that has been taken, the number of medicines, the medication time of each medicine, and the type and quantity of the remaining medicines according to the weight changes of all medicines, and generate a detailed report of the patient's medication and send it to the cloud data processing center and the patient's smart terminal APP through the wireless communication module. The central processing module is also used to control the alarm module to remind the medication time in real time. The central processing module is also used to determine whether the temperature, humidity and position balance exceed the threshold; the parameter setting module is used to set the threshold of the appropriate storage environment parameters of each medicine and the judgment threshold of the box being in a position balance state; the alarm module is used to automatically issue an alarm message when the temperature, humidity and position balance exceed the threshold or when the medication time of any medicine is reached. The micro-weight sensor uses an NFC smart sensor chip, which is encapsulated under each medicine position on the medicine board. The sensor in the chip detects the state change in real time and uploads it to the central processing module for processing. The box is used to store medicines for treating different chronic diseases. The micro-weight sensor is used to sense the changes in the environmental parameters in the box and the quantity and type of medicines in the box in real time, automatically identify the number of medicines on the medicine plate, record the time of medication for the patient, and generate a detailed medication report, and send the detailed medication report to the patient's smart terminal APP and the cloud data processing center. If the appropriate storage environment for the medicine is exceeded, a warning sound will be issued automatically. The position balance sensor 7 built into the box can monitor the state of the liquid medicine in real time to prevent it from spilling. Data collection is mainly completed by the intelligent sensor 9, the main components of which are: (1) temperature and humidity sensor 19, used to sense and record the temperature, humidity and other conditions of the environment where the medicine is located; (2) position balance sensor 7, used to sense whether the position of the medicine is horizontal (some liquid medicines need to be placed horizontally, otherwise they are easy to spill); (3) micro weight sensor, used to sense the weight of the medicine in the medicine box, used to record the time when the patient takes the medicine (when the sensor is attached to the bottom of the medicine plate, the number of tablets on the medicine plate is input in the corresponding mobile phone APP, and the sensor automatically calculates the weight of each tablet. When the patient takes the medicine, he knows how many tablets he has taken at a time. You can choose an NFC smart sensor chip, which can sense whether the medicine has been taken out and transmit data through NFC technology.The following are detailed implementation steps and sample code: Select the recommended NFC smart sensor chip, such as AMS AS39513, which is an NFC interface smart sensor tag chip with EEPROM storage and sensing functions. Encapsulate the chip under each drug position on the medicine board. Initialization and programming Use an NFC programming device (such as an NFC reader) to initialize each chip and write the drug information and initial status in the EEPROM. Real-time detection and data reading When the drug is taken out, the sensor in the chip detects the state change. The chip data is read through the NFC reader and uploaded to the central system for processing. ); (4) Light intensity sensor, used to sense and record the light intensity of the drug environment; (5) Switch controller: Record the time after the medicine box is opened and closed, and at the same time notify other sensors to record the temperature and humidity of the drug at that time, the amount of medicine, and other information. This device can collect data on the medication of patients with chronic diseases, such as medication time (accurate to seconds), medication type, and medication frequency.

[0035] Preferably, in another embodiment, the real-time management system for accurate medication of chronic disease comorbidity patients based on the AI ​​large model further includes a wearable module, which is a wearable adapter sensor, using a microneedle E-AB sensor array, inserting a tiny needle through the surface of the skin, extracting samples directly from sweat or blood, and real-time monitoring of the blood concentration of low-safety dose drugs. The blood concentration of low-safety dose drugs is transmitted to the patient's smart terminal APP via Bluetooth for real-time display, and uploaded to the cloud data processing center through the smart terminal APP, and then forwarded by the cloud data processing center to the physician smart terminal APP and the pharmacist smart terminal APP. Wearable module sensors based on microneedles are used for continuous real-time therapeutic drug concentration monitoring: microneedle E-AB sensor arrays are used for continuous monitoring of therapeutic drugs. Wearable blood drug concentration monitoring equipment uses microneedle array technology, inserting tiny needles through the surface of the skin, extracting samples directly from sweat or blood, and then using sensors for real-time monitoring. The sensor can detect drug concentrations as low as micrograms and transmit data to the user's smart device, such as a mobile phone or watch.

[0036] Then, the cloud data processing center processes and analyzes the data collected from the patient's smart terminal, smart sensor 9 and wearable module, which is mainly used in two aspects:

[0037] (1) Data recording and abnormal data distribution: The cloud data processing center uses this data to compare with the patient's current medication regimen. If any differences are found, such as reduced medication frequency or missed medications, an abnormal record is generated and synchronized to the physician and pharmacist for patient condition analysis. The patient is also notified through the APP to remind him to take the medication at a fixed time.

[0038] (2) Data application:

[0039] The AI ​​big model for safe medication guidance is built based on data from the following two sources: ① Medication data collected from patients with chronic diseases; ② Data composed of information such as patients' physical examination data in the hospital, past medical history, types of medications taken, etc., which are used to build the big model: The main data sources for building the big model: drug interaction databases such as Lexi-Interact, COVID-19, Paxlovid, Pharmapedia, CHDID, DDInter, Micromedex, DGIdb, BioGRID, DIDB, DDInter, Liverpool University HIV Drug Interaction Database, etc. and drug instructions for commonly used drugs for chronic diseases.

[0040] The AI ​​large model adopts a twin graph neural network algorithm based on path perception, generates edge feature tensors of the shortest reaction path distance between multiple path-aware drugs through multiple relay path sampling, and thus fuses the twin graph neural network to predict drug interactions. The construction method of the AI ​​large model includes: step S1: calling data in the drug interaction database, collecting drug instructions for commonly used drugs for chronic diseases, and medication data, physical examination data and medical history of several patients with chronic diseases; step S2: selecting the Transformer neural network structure, using the deep learning framework TensorFlow to build a neural network model, the model structure is: a large language model LLM composed of about 70k multi-head self-attention modules as the main modules, with a total of 50 billion parameters; cross-validation and grid search methods are used to optimize the model parameters. During the training process, the network weights are updated through the back-propagation algorithm, and the model performance is evaluated using the validation set, and the model parameters are continuously adjusted to achieve the optimal state.

[0041] Specifically, the construction process of the AI ​​big model for safe medication guidance is as follows:

[0042] 1) The main elements of input data include but are not limited to the generic name of the drug, dosage form, specifications, usage and dosage, contraindications, precautions for use, half-life, time to peak blood drug concentration, pharmacological effects, adverse reactions, drug interactions, use in special populations (such as pregnant and lactating women), pharmacokinetics, expiration date, etc.

[0043] 2) Select the Transformer neural network structure, use deep learning frameworks such as TensorFlow to build the neural network model, and optimize the model parameters by methods such as cross-validation and grid search. During the training process, the network weights are updated through the back-propagation algorithm, and the model performance is evaluated using the validation set, and the model parameters are continuously adjusted to achieve the optimal state. Model structure: The model LLM (Large Language Models) is composed of about 70k Multi-head Self-Attention modules as the main modules, with a total of 50 billion parameters. The LLM model has three main parts: ① Input Encoder: Token Embedding: Map each word or subword (token) in the input text to a high-dimensional vector representation, usually using pre-trained word embedding (such as Word2Vec, GloVe); Positional Encoding: In order to retain the order information of the input sequence, position encoding is added to the embedding vector of each word, usually using sine and cosine functions for encoding. ②Multi-head Self-Attention: Allows the model to pay attention to other positions in the input sequence at different positions to capture global dependencies. ③Output layer: The output layer is usually a fully connected layer that outputs the text of the last layer of the Transformer Encoder. LLM neuron diagram is shown below Figure 2 shown.

[0044] 3) Output data is the target value of the model, including the usage and dosage of each drug, precautions, dietary taboos, drug interaction risk warnings, lifestyle guidance, etc. For example, for drug usage, the big model recommends patients to take the drug in the morning or at night based on the drug's pharmacological mechanism, half-life and other information; for drug dosage, the big model recommends the best dosage based on the patient's basic information of height and weight, realizes personalized medication, and reduces adverse drug reactions; for potential drug interaction risk warnings, the big model reminds patients of possible adverse reactions or replaces other suitable drugs based on the drug's pharmacokinetic characteristics, whether there is liver drug enzyme induction, whether there is a competitive metabolic pathway, etc.; for drug contraindications, such as prohibiting the use of alcoholic beverages when taking cephalosporins, etc. According to the patient's medical history and physical examination results, the big model can also give some lifestyle suggestions and give patients improvement measures that can be taken in daily life, such as dietary advice, exercise guidance, psychological support, etc.

[0045] Based on the patient's basic information, physical examination data, past medical history and all the information of the medicines being taken, the cloud data processing center uses the AI ​​big model to analyze the potential risks of drug interactions, the precautions for the use of various drugs, and gives dietary and lifestyle recommendations. Finally, safe medication guidance information is generated and sent to the patient's smart terminal APP in real time to guide the patient's safe medication. The safe medication guidance information includes real-time medication reminders, drug interaction risk reminders, medication precautions, dietary taboos and lifestyle guidance, so as to guide the patient to use medication safely; it is also used to forward the patient's medication details report uploaded by the micro-weight sensor to the physician's smart terminal APP and the pharmacist's smart terminal APP in real time. The physician APP, pharmacist APP, physician PC client, pharmacist PC client, and patient APP share data information through the cloud data processing center to achieve accurate tracking and management of patient medication by physicians and pharmacists. Through big models, artificial intelligence neural network algorithms, deep learning and other technologies, the patient APP can receive information such as medication time, drug interaction risk reminders, medication precautions, dietary taboos and lifestyle guidance to guide patients to use medication safely.

[0046] like Figure 3-Figure 7As shown, the medicine box 1 includes a box body, an inner frame 2 and a cover plate 3, the cover plate 3 is connected to the upper part of the box body by a hinge, the inner lower part of the box body is provided with a bottom plate 17, the bottom plate 17 is provided with a tablet plate 4 and a liquid medicine placement plate 5, the tablet plate 4 is provided with a micro-weight sensor (which can automatically detect the number of medicines on the plate by magnetic adsorption, automatically record the time of taking each medicine, and upload it to the cloud and the patient's smart terminal in real time), the tablet plate 4 is provided with a tablet placement area 8, the liquid medicine placement plate 5 is provided with a liquid medicine bottle placement area 6, and a partition plate 10 is provided between the tablet plate 4 and the liquid medicine placement plate 5. The liquid medicine placement plate 5 is embedded with a position balance sensor 7, which can monitor the state of the liquid medicine in real time to prevent it from spilling, and the cover plate 3 is embedded with a detection device 26 composed of a temperature sensor 18, a humidity sensor 19 and a light intensity sensor 20. The temperature sensor 18 is connected to a temperature display device 22, the humidity sensor 19 is connected to a humidity display device 23, the light intensity sensor 20 is connected to a light intensity display device 24, and the temperature sensor 18 and the humidity sensor 19 are both located inside the medicine box 1. The detection device 26 is provided with a detection head 21, and the detection head 21 is connected to an alarm module 25. The detection device 26 is located on the upper part of the tablet medicine board 4 and the liquid medicine placement board 5, and a partition plate 10 is provided between the tablet medicine board 4 and the liquid medicine placement board 5. Connecting members 11 are provided on both sides and the lower part of the partition plate 10 on the inner frame 2, and a limiting shaft 12 is provided on the connecting member 11. Limiting members 13 are provided on both sides and the lower part of the partition plate 10, and the limiting shaft 12 runs through the interior of the limiting member 13. The inner side of the inner frame 2 is also provided with a mounting member 14 at the lower part of the partition plate 10, and the mounting member 14 A screw rod 15 is rotatably installed on the inner side, and the screw rod 15 is threadedly connected to the middle part of the lower side of the partition plate 10. One end of the screw rod 15 is connected to a drive motor 16. The output end of the drive motor 16 is connected to one end of the screw rod 15 through a coupling. The drive motor 16 is connected to the inner frame 2. A movable groove 27 is opened at the lower part of the bottom plate 17. The lower part of the partition plate 10 and the limiting member 13 located at the lower part of the partition plate 10 are located inside the movable groove 27, and the width of the movable groove 27 is matched with the width of the limiting member 13.

[0047] The embodiments of the present invention are given for the purpose of illustration and description, and are not intended to be exhaustive or to limit the invention to the disclosed forms. Many modifications and variations will be apparent to those of ordinary skill in the art. The embodiments are selected and described in order to better illustrate the principles and practical applications of the present invention and to enable those of ordinary skill in the art to understand the present invention and thereby design various embodiments with various modifications suitable for specific uses.

Claims

1. A real-time management system for accurate medication of patients with chronic diseases based on AI big model, characterized by: include: Smart medicine box, patient smart terminal APP, physician smart terminal APP, pharmacist smart terminal APP and cloud data processing center; The smart medicine box includes a medicine box and a smart sensor arranged in the medicine box. The medicine box is used to store medicines for treating different chronic diseases. The smart sensor is used to sense the changes in the environmental parameters in the box and the quantity and type of medicines in the box in real time, and generate a medication details report, and send the medication details report to the patient smart terminal APP and the cloud data processing center; The patient smart terminal APP is used to set the medication time for different medicines; It is also used to receive the detailed medication report sent by the smart sensor; it is also used to obtain the patient's basic information, physical examination data, past medical history and all medication information currently being taken, and upload the above information to the cloud data processing center; The cloud data processing center is used to analyze potential drug interaction risks, medication precautions of various drugs and give diet and lifestyle recommendations based on the patient's basic information, physical examination data, past medical history and all medication information currently being taken, using an AI large model, and ultimately generates safe medication guidance information and sends it to the patient's smart terminal APP in real time to guide the patient's safe medication use. The safe medication guidance information includes real-time medication reminders, drug interaction risk reminders, medication precautions, dietary taboos and lifestyle guidance; it is also used to forward the patient's medication details report uploaded by the smart sensor to the physician's smart terminal APP and the pharmacist's smart terminal APP in real time.

2. A real-time management system for accurate medication for patients with chronic comorbidities based on an AI big model as claimed in claim 1, characterized in that: The AI ​​big model adopts a twin graph neural network algorithm based on path perception, and generates edge feature tensors of the shortest reaction path distances between multiple path-aware drugs through multiple relay path sampling, thereby fusing the twin graph neural network to predict drug interactions.

3. A real-time management system for accurate medication for patients with chronic comorbidities based on an AI big model as claimed in claim 2, characterized in that: The method for constructing the AI ​​big model includes: Step S1: calling the data in the drug interaction database, collecting the drug instructions of commonly used drugs for chronic diseases and the medication data, physical examination data and medical history of several patients with chronic diseases; Step S2: Select the Transformer neural network structure and use the deep learning framework TensorFlow to build a neural network model. The model structure is: a large language model LLM composed of approximately 70k multi-head self-attention modules as the main modules, with a total of 50 billion parameters; use cross-validation and grid search methods to optimize the model parameters. During the training process, the network weights are updated through the back-propagation algorithm, and the model performance is evaluated using the validation set, and the model parameters are continuously adjusted to achieve the optimal state.

4. A real-time management system for accurate medication for patients with chronic comorbidities based on an AI big model as claimed in claim 1, characterized in that: The invention further includes a wearable module, which adopts a microneedle E-AB sensor array, inserts tiny needles through the surface of the skin, directly extracts samples from sweat or blood, and monitors the blood concentration of low-safety-dose drugs in real time. The blood concentration of low-safety-dose drugs is transmitted to the patient's smart terminal APP via Bluetooth for real-time display, and is uploaded to the cloud data processing center through the smart terminal APP, and then forwarded by the cloud data processing center to the physician smart terminal APP and the pharmacist smart terminal APP.

5. A real-time management system for accurate medication for patients with chronic comorbidities based on an AI big model as claimed in claim 1, characterized in that: The intelligent sensor includes a humidity sensor, a temperature sensor, a position balance sensor, a micro-weight sensor, a light intensity sensor, a central processing module, a timer, a parameter setting module, an alarm module, a power module, a wireless communication module and a switch controller; the switch controller is used to record the time after the medicine box is opened and closed, and notify other sensors to record the environmental parameters and drug information of the medicine at that time; the central processing module is used to automatically identify the type of medicine that has been taken, the number of medicines, the medication time of each medicine, and the type and quantity of the remaining medicines according to the weight changes of all medicines, and generate a detailed report on the patient's medication and send it to the cloud data processing center and the patient's smart terminal APP through the wireless communication module. The central processing module is also used to control the alarm module to remind the medication time in real time, and the central processing module is also used to determine whether the temperature, humidity and position balance exceed the threshold; the parameter setting module is used to set the threshold of the appropriate storage environment parameters of each medicine and the judgment threshold of the box being in a position balance state; the alarm module is used to automatically issue an alarm message when the temperature, humidity and position balance exceed the threshold or when the medication time of any medicine is reached.

6. A real-time management system for accurate medication for patients with chronic comorbidities based on an AI big model as claimed in claim 5, characterized in that: The micro-weight sensor adopts an NFC smart sensor chip, which is encapsulated under each medicine position on the medicine plate. The sensor in the chip detects the state change in real time and uploads it to the central processing module for processing.

7. A real-time management system for accurate medication of chronic disease comorbidity patients based on AI big model as claimed in claim 1, characterized in that: The medicine box includes a box body, an inner frame and a cover plate, the cover plate is connected to the upper part of the box body by a hinge, the inner lower part of the box body is provided with a bottom plate, a tablet medicine board and a liquid medicine placement board are provided on the bottom plate, a micro-weight sensor is provided on the tablet medicine board, a position balance sensor is embedded on the liquid medicine placement board, a detection device composed of a temperature sensor, a humidity sensor and a light intensity sensor is embedded on the cover plate, the detection device is located on the upper part of the tablet medicine board and the liquid medicine placement board, a partition plate is provided between the tablet medicine board and the liquid medicine placement board, and connectors are provided on both sides and the lower part of the inner frame at the partition plate, the connector A limiting shaft is arranged on the top, limiting parts are arranged on both sides and the lower part of the partition plate, the limiting shaft passes through the interior of the limiting part, and a mounting part is also arranged on the inner side of the inner frame at the lower part of the partition plate, a screw is rotatably installed on the inner side of the mounting part, the screw is threadedly connected to the middle part of the lower side of the partition plate, one end of the screw is connected to a driving motor, the output end of the driving motor is connected to one end of the screw through a coupling, the driving motor is connected to the inner frame, a movable groove is opened at the lower part of the bottom plate, the lower part of the partition plate and the limiting part located at the lower part of the partition plate are located inside the movable groove, and the width of the movable groove is matched with the width of the limiting part.

8. A real-time management system for accurate medication for patients with chronic comorbidities based on an AI big model as described in any one of claims 1 to 7, characterized in that: The intelligent terminal includes a mobile terminal and a PC terminal.

Images