Intelligent medicine box based on patient discharge medication

Abstract

The invention provides an intelligent medicine box based on patient discharge medication. The intelligent medicine box comprises a hardware body and a matched control module. The matched control module comprises a remote information input sub-module, a data storage sub-module, a trigger control sub-module, a medicine approval sub-module and an intelligent analysis and decision-making sub-module; the remote information input sub-module obtains data related to diagnosis and treatment of a patient from a hospital end; the data storage sub-module collects and stores various data; the trigger control sub-module triggers the corresponding multimedia elements based on the signals returned by the elements of the various sensing units; the intelligent analysis and decision submodule formulates a personalized medication strategy based on the data from the hospital end and the physiological life data of the patient. Through the medicine box, the effect of guarding the medication safety of a patient after the patient is discharged from a hospital is achieved.

Description

Technical Field

[0001] This invention relates to the field of intelligent medical devices and medication education guidance technology. Specifically, it is an intelligent pillbox for discharged patients that has the ability to self-build patient models, make dynamic decisions on multi-dimensional risks, and self-optimize intervention strategies. It is used to solve the problems of discharged patients forgetting pharmacist medication instructions and being confused about drug information. It can automatically synchronize drug information and educational content when taking medication, thereby improving medication safety and adherence. Background Technology

[0002] Discharged patients are a key group for home medication. Their medication regimens are usually educated and guided by hospital pharmacists at discharge through verbal explanations, written instructions, or single-page manuals. However, in practice, there are significant pain points as follows:

[0003] 1. Information forgetting and passive recall: After being discharged from the hospital, patients are affected by memory decline and are prone to forgetting key information such as medication contraindications and adverse reaction management emphasized by the pharmacist. Generally, patients need to actively refer to the medical information or drug instructions to confirm the use of the drug, which is extremely inconvenient in medication scenarios.

[0004] 2. Disconnect between static reminders and dynamic behavior: Most smart pillboxes on the market only have fixed time-based reminder functions, which cannot detect whether the patient has actually taken the medicine, let alone determine whether the time and dosage of the medicine are correct, and lack effective intervention for missed or incorrect doses.

[0005] 3. Medication guidance is decoupled from physiological state: Existing protocols cannot obtain real-time physiological state (such as blood pressure and heart rate) of patients when taking medication, which makes it impossible to stop high-risk medication behaviors such as taking antihypertensive drugs when blood pressure is too low.

[0006] 4. Information silos: As an independent device, the pillbox cannot be linked with the increasingly popular wearable health devices in the home, thus failing to form an integrated health management loop.

[0007] Therefore, there is an urgent need for a smart pillbox that can integrate medication information management, behavior perception, physiological data fusion and intelligent decision-making intervention to fill the gap in medication education for discharged patients at home and achieve a leap from "passive reminders" to "proactive protection". Summary of the Invention

[0008] This invention aims to overcome the aforementioned shortcomings and provide a novel intelligent medicine box for patient discharge medication education. By introducing a self-construction mechanism for personalized patient models and a dynamic optimization mechanism for multi-dimensional risk decision-making and intervention strategies, the system possesses the capabilities of autonomous initialization, continuous learning, and personalized adaptation. This completely eliminates reliance on manually preset rules, achieving a leap from "static execution" to "dynamic intelligence." The specific solution is as follows:

[0009] This invention provides an intelligent medicine box based on patient discharge medication, characterized in that it includes a hardware body and a supporting control module;

[0010] The aforementioned hardware unit includes a drug storage unit, an information interaction unit, and a triggering and sensing unit.

[0011] The aforementioned medicine storage unit contains several medicine compartments;

[0012] The aforementioned information interaction unit includes multimedia components;

[0013] The aforementioned sensing unit includes a biometric element for opening the medicine box lid, a gravity sensor for identifying the weight of the medicine in the medicine compartment, and a lid opening sensor for detecting whether the medicine compartment lid is open or not.

[0014] The electrical control components within the aforementioned hardware are all electrically connected to the corresponding control module;

[0015] The aforementioned supporting control modules include a remote information input submodule, a data storage submodule, a trigger control submodule, a drug approval submodule, and an intelligent analysis and decision-making submodule;

[0016] The aforementioned remote information entry submodule acquires data related to patient diagnosis and treatment from the hospital.

[0017] The aforementioned data storage submodule collects and stores various types of data;

[0018] The aforementioned trigger control submodule triggers the corresponding multimedia elements based on the signals transmitted back from the elements of various sensing units.

[0019] The aforementioned intelligent analysis and decision-making submodule formulates personalized medication strategies based on data from the hospital and patients' physiological and lifestyle data.

[0020] Furthermore, the present invention provides an intelligent pillbox based on patient discharge medication, characterized in that:

[0021] The aforementioned trigger control submodule,

[0022] Personalized information is displayed based on the differences between the people who open the medicine box;

[0023] Personalized information is displayed based on the differences in the opened medicine compartments;

[0024] Based on the information transmitted back from the sensors, an alert is issued when the number or type of medication taken by the patient is incorrect.

[0025] Furthermore, the present invention provides an intelligent pillbox based on patient discharge medication, characterized in that:

[0026] The aforementioned intelligent analysis and decision-making submodule includes a patient-personalized strategy self-construction unit, a multi-dimensional risk decision engine, and an AI question-answering unit. The optimal medication strategy is formed through the integration of these three components.

[0027] Furthermore, the present invention provides an intelligent pillbox based on patient discharge medication, characterized in that:

[0028] The aforementioned patient-personalized model self-construction unit calls upon diagnosis and treatment-related information, and formulates / adjusts / modifies the patient's medication plan based on medication rules and the patient's own lifestyle and physiological adaptability.

[0029] Furthermore, the present invention provides an intelligent pillbox based on patient discharge medication, characterized in that:

[0030] The aforementioned multi-dimensional risk decision engine identifies whether a patient is in a specific environment or scenario by monitoring the patient's physiological data, thereby determining whether the patient needs to adjust their medication strategy.

[0031] The specific method is as follows:

[0032] S1. Receive and integrate real-time physiological data, the risk attributes of the drug itself, the patient's personalized historical records, and related real-time and historical data;

[0033] S2. Use weighted logistic regression or lightweight neural network models to comprehensively calculate the above multidimensional features and distinguish between physiological fluctuations and pathological abnormalities;

[0034] S3. Connect the built-in clock and weight sensor to compare the actual medication time and amount with the preset medication plan to determine the impact of missed doses, wrong time of administration, and similar behaviors on these physiological data;

[0035] S4. Based on the above-mentioned influencing relationships, formulate dynamic intervention strategies according to different scenarios and levels.

[0036] Furthermore, the present invention provides an intelligent pillbox based on patient discharge medication, characterized in that:

[0037] The aforementioned multi-dimensional risk decision engine analyzes the medical information sent by the hospital during the patient's medication period to determine whether or not the medication strategy needs to be adjusted.

[0038] The specific method is as follows:

[0039] S1. Receive medical information from the hospital;

[0040] S2. Determine whether the patient’s current medication conflicts with any recent medical events that are in progress, have been, or are about to occur;

[0041] S3. If a conflict exists, adjust the medication regimen.

[0042] Furthermore, the present invention provides an intelligent pillbox based on patient discharge medication, characterized in that:

[0043] The aforementioned AI question-answering unit provides patients with professional knowledge related to medication, captures abnormalities that cannot be detected by other physiological monitoring devices during medication use, and incorporates them into the formulation of medication strategies.

[0044] Furthermore, the present invention provides an intelligent pillbox based on patient discharge medication, characterized in that:

[0045] It also includes a drug approval submodule;

[0046] The aforementioned drug approval submodule compiles statistics on drug warehousing during the drug placement stage and drug warehousing during the drug retrieval stage.

[0047] Furthermore, the present invention provides an intelligent pillbox based on patient discharge medication, characterized in that:

[0048] It also includes a risk warning module;

[0049] The aforementioned risk warning module sends alert messages to patients / patients' families / hospital doctors when an abnormal event occurs. Attached Figure Description

[0050] Figure 1 Schematic diagram of the medicine box structure;

[0051] Figure 2 System logic diagram;

[0052] Figure 3 A schematic diagram of a display interface. Detailed Implementation

[0053] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0054] This embodiment provides a smart pillbox, which includes a hardware body and a supporting control module;

[0055] Multiple components within the hardware unit are electrically connected to the corresponding control module;

[0056] The hardware itself, such as Figure 1 As shown, it includes:

[0057] Medication storage unit 1: Divided into several independent medication storage compartments according to the types of medications patients need upon discharge (e.g., 1-10 detachable / fixed compartments). Each compartment is marked with a unique "compartment number" on the outside and has an RFID tag on the inside.

[0058] Information interaction unit: includes various multimedia components, such as a touch screen 3 (≥3.5 inches) on the front of the medicine box, a voice broadcaster 6, and an integrated microphone 4, thereby supporting voice interaction function.

[0059] The voice broadcaster and integrated microphone can automatically provide audio reminders at specified times according to medication rules. The microphone can also recognize user commands through voice.

[0060] Power supply unit: A rechargeable lithium battery (with a battery life of ≥7 days) is built into the entire pillbox and can be charged via the equipped Type-C charging port 5.

[0061] Triggering and sensing units: These are mainly involved in the opening and dispensing of medications, specifically including:

[0062] A. Biometric Module: An integrated fingerprint reader 7 is used to verify the user's identity. When the user is a preset user (i.e., the patient, a family member assisting with filling in medication / entering information, or a medical staff member), the medicine box lid 2 can be opened by fingerprint. The lid is generally made of transparent material, so that the information displayed on the touch screen can be seen directly from the outside. The integrated fingerprint reader 7 can also enable the matching control module to determine the identity of the current operator through fingerprint recognition, thereby pushing different display information. For example, when the operator is the patient, the information of the medication to be taken can be displayed. When the operator is someone else, the patient's medication status and information on medication supplementation or adjustment can be displayed.

[0063] B. Weight sensor: Integrated at the bottom of each medicine storage compartment, used to accurately monitor the weight change after each medicine dispensing, thereby determining whether medicine should be dispensed and whether the amount dispensed is accurate.

[0064] C. Each medicine storage compartment is equipped with a cover containing an opening sensor (such as a Hall sensor) to detect the opening action.

[0065] The supporting control module includes: a remote information input submodule, a data storage submodule, a trigger control submodule, a drug approval submodule, and an intelligent analysis and decision-making submodule;

[0066] The modules have an information exchange and linkage effect;

[0067] Among them, the remote information entry submodule supports connection to hospital workstations or dedicated physician apps via Bluetooth / WiFi, enabling pharmacists to enter basic drug information and customized medication information based on patient signs on their work terminals (e.g., pharmacist terminals). After collecting the information transmitted in this way, this module stores it in the data storage submodule of this medicine box.

[0068] Data storage submodule: Built-in local storage chip (capacity ≥16GB) for encrypted storage of patient information, medication strategies, targeted drug education / teaching content, customized supporting information, and historical medication data. It also stores information transmitted from various external devices, such as hospital / physician records, diagnoses, and medical orders, as well as information transmitted from wearable devices (mainly those related to physiological data monitoring).

[0069] Trigger control submodule: Based on feedback signals from various sensors, it triggers corresponding multimedia components. For example, it provides personalized prompts for the person opening the medicine box, displays personalized information about the opened medicine compartment, and provides voice announcements. Specifically:

[0070] (1) Regarding the medicine box itself: When the medicine box is opened based on biometrics, the trigger control submodule retrieves the identity information of the person opening the box from the data storage submodule and displays the corresponding prompt information for that person on the display screen. For example, if the person opening the box is the patient, the screen displays the medication information that needs to be taken at the current time, including: medication dosage, medication method (e.g., before meals, after meals, with meals, diluted with water according to the ratio, etc.). If there is no medication that needs to be taken at the current time, the screen displays the status of the medication to be taken, the status of the medication already taken, and whether medication needs to be replenished.

[0071] In some experimental cases, the lid of the medication compartment can also be automatically opened. If the person opening the lid is a doctor or family member, the screen will display information about the patient's medication, such as whether the medication was taken on time, whether a dose was missed, and whether the medication needs to be replenished.

[0072] (2) For medicine compartment: When the module receives the feedback signal that the opening sensor is activated, it matches and displays the corresponding medicine information and specific precautions on the screen for the medicine corresponding to the opened medicine compartment.

[0073] (3) When the medication is due for administration, if the sensor detects that the patient has taken the wrong amount or type of medication, a warning message will be displayed on the screen and a voice prompt will be triggered.

[0074] Drug approval submodule: This module tracks drug warehousing during the drug placement phase and drug retrieval phase. Specifically:

[0075] (1) For drug warehousing: This includes two cases: initial warehousing and intermediate warehousing. It is linked with the trigger control submodule. When the medicine box is opened and the person opening the box selects the drug warehousing option, the screen displays the storage status of each medicine compartment, in a format such as 1 / 10, where 1 is the remaining quantity and 10 is the initial warehousing quantity; if it is the first warehousing, it displays 0 / 0. After the person opening the box opens the medicine compartment one by one, the interface displays the drug information of the medicine compartment. If it is the first warehousing or the current inventory is 0, the interface for information to be entered is displayed, and the person opening the box needs to actively enter the drug information (or select the drug based on the patient's prescription stored in the database). After the person opening the box completes the warehousing, they also need to enter the quantity information of the drug, so that the module will actively calculate the average weight of a single drug. After all the drugs are warehousing, save and exit the interface to display the status of all warehousing compartments.

[0076] (2) For medication dispensing: This includes two scenarios: patient medication collection and medication adjustment by related persons / patients. It is linked to the trigger control submodule. When the medication compartment is opened, the module records the weight loss from opening to closing each time in real time. For example, if the person opening the medication compartment is a patient and it is opened within the medication time, the weight loss will be recorded as part of the medication administration record. If the weight loss exceeds the policy's limits, a warning message will be issued on the screen and voice control devices until the patient corrects the operation or selects the medication replacement / cancellation option on the screen. If the person opening the compartment is someone else or a patient and it is not within the medication time, the screen will prompt whether to use the medication replacement interface. After selecting this interface, the user can remove all the medication without warning. If there is any residue, the screen and voice control will also issue a warning.

[0077] The intelligent analysis and decision-making submodule (the main calculation process can exist on remote devices such as servers / APPs, while the main part of this medicine box is to display, receive and transmit data) includes: a patient personalized strategy self-construction unit, a multi-dimensional risk decision engine and an AI question-answering unit, which form the optimal medication strategy through the integration of the three.

[0078] The patient-personalized model self-construction unit calls upon information from the database related to diagnosis and treatment provided by hospitals / physicians. Based on medication rules and the patient's lifestyle and physiological adaptability [including general monitoring results and patient-reported adverse reactions (this content is obtained from the AI ​​Q&A process)], it formulates / adjusts / modifies the patient's medication plan. The specific methods are as follows:

[0079] S1. Call the patient information received from the hospital data terminal in the data storage submodule and import it into this module. Analyze the patient's discharge medical orders, medication regimens, allergy history and underlying disease information, and autonomously generate an initial personalized medication rule base, greatly reducing the burden of manual data entry and the error rate.

[0080] S2. During the system initialization phase (e.g., the first two weeks), the system continuously records the patient's actual medication collection behavior, automatically identifies the patient's personalized medication time pattern through clustering algorithms, and dynamically corrects the preset "standard medication time window" in the system accordingly, forming a personalized medication time window that fits the patient's daily routine.

[0081] The S3 system, linked to wearable devices, acquires information returned by the patient's wearable device. Within a specified period (e.g., 7 consecutive days), it collects the patient's physiological data at rest. Through statistical analysis, it automatically calculates and establishes the patient's personal physiological safety baseline (e.g., personal resting blood pressure range) to replace universal fixed thresholds, achieving individualized safety assessments. Based on the patient's physiological information, it adjusts medication timing to obtain the optimal dosage. For example, by monitoring the patient's physiological data, it can discover patterns in physiological data changes and medication use (this process generally requires a certain period). When certain physiological changes may affect the patient's health, adjusting the medication timing can be experimentally used to see if these changes can be avoided or mitigated.

[0082] Multi-dimensional risk decision engine: By monitoring the patient's physiological data, it identifies whether the patient is in a specific environment or scenario, thereby determining whether the patient can take medication according to the usual strategy; the specific methods are as follows:

[0083] S1. Receive and integrate real-time physiological data (such as current blood pressure, blood sugar, heart rate, etc.), the risk attributes of the drug itself (risk level extracted from the pharmaceutical knowledge base), and the patient's personalized historical records (such as the frequency of recent abnormal events and personal physiological baseline), as well as other real-time and historical data.

[0084] S2. Use weighted logistic regression or lightweight neural network models to comprehensively calculate the above multidimensional features and distinguish between physiological fluctuations and pathological abnormalities;

[0085] S3. Connect the built-in clock and weight sensor to compare the actual medication time and amount with the preset medication plan to determine the impact of missed doses, wrong time of administration, and other behaviors on these physiological data;

[0086] S4. Based on the above-mentioned influencing relationships, formulate dynamic intervention strategies based on different scenarios and levels. For example, when physiological monitoring indicates that the patient may be exercising, for the scenario of "taking antihypertensive drugs after exercise," the output strategy is "prompt rest and retest" instead of "force lockout." If the retest result meets the standard, the current medication strategy is considered OK. If the retest result is still unsatisfactory, the strategy is adjusted to "postpone medication" or "reduce the dosage in the current time period," etc. By dynamically adjusting medication and combining it with the changing patterns of physiological data, the effect of optimizing the medication strategy can be achieved.

[0087] In addition, by analyzing the patient's medical information during medication sent by the hospital, it is possible to determine whether or not the medication strategy needs to be adjusted. The specific methods are as follows:

[0088] S1. Receive medical information from the hospital;

[0089] S2. Determine whether the patient’s current medication conflicts with any recent medical events that are in progress, have been, or are about to occur;

[0090] S3. If a conflict exists, adjust the medication regimen.

[0091] For example, if a patient needs to undergo surgery or other medical procedures that explicitly prohibit the use of anticoagulants, and the patient already has anticoagulants in their regular medications, the supply of that medication will be suspended when the medicine box detects this information.

[0092] The AI-powered question-answering unit: On the one hand, it provides patients with professional knowledge related to medication; on the other hand, it captures abnormalities that occur during medication use that cannot be detected by other physiological monitoring devices, and incorporates these abnormalities into the formulation of medication strategies. Specifically:

[0093] (1) By connecting to the cloud-based pharmaceutical knowledge base via Wi-Fi, patients can input their medication questions through a microphone or touchscreen, and the system provides instant, professional answers in voice or text format. Based on the general pharmaceutical knowledge base, it deeply integrates personal data from the "patient personalized model self-construction module". Its answer engine is configured to: link patient questions with personal medication plans, historical physiological data, allergy history, and recent abnormal records in real time to generate highly personalized medication recommendations.

[0094] (2) When a patient raises questions about special reactions after medication through this Q&A method, this unit captures the patient's self-reported special reactions and introduces them into the personalized model self-construction unit, making them one of the references for strategy formation. For example, if a patient reports symptoms such as abdominal distension and dizziness after taking medication, the AI ​​unit identifies the patient's information, searches whether the patient's medication has such adverse reactions, and if so, provides verbal education about the situation. At the same time, the system reports the patient's current adverse reactions to the system, which records the information and also feeds it back to the hospital. If there is an alternative medication, the hospital can remotely change the medication, thus achieving a positive linkage effect. Similarly, if there is a medication change scenario, the historical data is recorded, which serves as one of the bases for analyzing the impact of the medication on the patient. Through monitoring, it is possible to determine whether the patient is suitable to use drug A or drug B, thereby achieving personalized customization. Of course, there are other scenarios as well. The core of this AI unit is to further enrich the collection of patient adverse reactions and improve the linkage effect between hospitals and patients.

[0095] In addition, as a supporting control module, a risk warning module is also included. This module is equipped with a reporting mechanism that links with the patient / family / hospital doctor's end. That is, when an abnormal event occurs, a warning message is sent to the patient / family / hospital doctor's end. For example, in the form of an app, patients / family members can understand the patient's medication status in real time (all medication-related matters, including medication frequency, whether medication is missed or taken incorrectly, changes in physiological data, etc.), the status of missing or needing to be replenished in the medicine box, and the hospital's adjustment strategies / notifications, etc.

[0096] The following is an application example of using the above-mentioned pillbox:

[0097] A 65-year-old diabetic patient with hypertension needs to take metformin tablets and nifedipine controlled-release tablets upon discharge.

[0098] The process for achieving enhanced personalized medication based on this smart pillbox is as follows:

[0099] A1. Pharmacist information entry stage upon discharge:

[0100] The pharmacist connects the smart pillbox to the hospital tablet (via Bluetooth) and, in the "Pharmacist Information Input Submodule," binds medications to each of the pillbox's two storage compartments.

[0101] Cabin 1: Basic drug information (Name: Metformin tablets, Specification: 0.5g / tablet, Dosage and administration: 1 tablet 3 times a day); Customized educational information ("Take 30 minutes after meals. If diarrhea occurs, observe for 3 days. If it does not improve, contact an endocrinologist. Avoid alcohol.")

[0102] Chamber 2: Basic drug information (name: nifedipine controlled-release tablets, specification: 30mg / tablet, dosage: 1 tablet once daily); Customized educational information ("Swallow the whole tablet before breakfast, do not break or chew it, monitor blood pressure, if it is below 110 / 70mmHg, stop taking the medication and contact the cardiology department").

[0103] After the data entry is completed, the pharmacist previews the information on the medicine box's touchscreen, and after confirming that everything is correct, disconnects the connection and delivers the medicine box and the medication (which has been dispensed into the corresponding compartment) to the patient.

[0104] 2. The patient's home medication stage:

[0105] Physiological baseline calibration: In the following week, the medication habit learning unit found that the patient always took nifedipine around 8:15, so the "personalized medication time window" was adjusted to 8:05-8:35. The physiological baseline calibration unit simultaneously calculated the patient's morning resting blood pressure baseline to be 115 / 75 mmHg.

[0106] Scenario 1: Individualized Risk Assessment

[0107] One morning, after exercising (8:20 AM), the patient prepared to take medication. Their real-time blood pressure was 105 / 65 mmHg. A multi-dimensional risk decision engine analysis revealed: ① While the current blood pressure (105 / 65) was below the general threshold, it was close to the individual's baseline (115 / 75); ② The medication risk was high; the wearable device engine determined this to be "physiological fluctuation after exercise," with a moderate risk score. Dynamic intervention: The system displayed: "We detected that you have just finished exercising, and your blood pressure is temporarily low. We recommend resting for 15 minutes until your blood pressure returns to normal before taking your medication."

[0108] Scenario 2: Intervention for Missed Administers

[0109] One day after lunch, the patient forgot to take metformin. Within the preset time window, no weight change was detected in compartment 1. The compliance analysis unit of the intelligent analysis and decision-making submodule determined it as "missed dose".

[0110] When it's time to take the medication after dinner, the pillbox not only provides the usual reminder, but the touchscreen also displays: "The system has detected that you may have missed your metformin dose after lunch. Please take the usual dose this time, and do not double it." A voice announcer simultaneously broadcasts this message.

[0111] Scenario 3: Staggered Dosing and Safety Warnings

[0112] The patient felt dizzy before bed and decided to take an extra dose of antihypertensive medication. When they opened the No. 2 medication compartment (nifedipine controlled-release tablets), both the opening sensor and the weight sensor were triggered simultaneously. The compliance analysis unit first determined that this was an unplanned medication time, and the intervention strategy engine immediately instructed the touch screen to highlight a warning: "Warning! This is an unplanned medication time. Please confirm if this is a new doctor's order. Do not take the medication unless necessary."

[0113] Meanwhile, the safety verification unit obtains the real-time blood pressure value of 85 / 55 mmHg from the smart bracelet worn by the patient via Bluetooth.

[0114] The intervention strategy engine combines two high-risk signals, "mistimed" and "low blood pressure," to trigger the highest level alarm: the screen turns red and displays: "Emergency alarm! Your blood pressure is detected to be too low. Taking antihypertensive medication poses a serious risk! We strongly recommend that you measure your blood pressure immediately and contact your doctor!" At the same time, the medicine box continuously emits an alarm sound and locks the medicine compartment until the patient presses and holds the confirmation button to acknowledge the risk.

[0115] Scenario 4: Intelligent Question Answering

[0116] After taking metformin, the patient remembered he had a party that evening and asked the pillbox, "Can I drink alcohol?"

[0117] The AI ​​Q&A submodule was activated and answered via voice: "According to your medication record, you should avoid drinking alcohol while taking metformin. Alcohol may increase the risk of lactic acidosis and may mask symptoms of hypoglycemia. We recommend that you drink tea instead of alcohol."

[0118] The function and effect of this embodiment:

[0119] This embodiment overcomes the passive response of traditional medicine boxes and achieves proactive intervention: through behavioral perception (weight + time), it realizes accurate judgment and real-time intervention of medication adherence, effectively reducing the rate of missed doses and incorrect doses.

[0120] This embodiment establishes a contextualized safety barrier. By linking with wearable devices, it adds a safety check based on physiological state before medication, avoiding the risks that may be caused by medication in abnormal states.

[0121] This embodiment realizes closed-loop medication management, forming a complete management closed loop of "planning-reminder-behavior perception-safety verification-intervention-recording-repurchase reminder".

[0122] This embodiment realizes an intelligent interactive experience, integrates AI Q&A, solves the randomness problem in the patient's medication process, effectively discovers adverse reactions that cannot be captured by traditional monitoring devices during the patient's medication, and upgrades the medicine box from a tool to a "personal pharmacist".

[0123] This embodiment ensures and enhances safety by using biometrics to effectively prevent accidental ingestion by children or others, thus improving family medication safety.

Claims

1. An intelligent medicine box based on patient discharge medication, characterized in that, The hardware body and the matching control module are included. The hardware body includes a medicine storage unit, an information interaction unit, and a triggering and sensing unit. The medicine storage unit includes a plurality of medicine cabins. The information interaction unit includes a multimedia element. The sensing unit includes a biometric element for opening the medicine box cover, a gravity sensor for identifying the weight of the medicine in the medicine cabin, and an opening sensing sensor for detecting whether the medicine cabin cover is opened. The electrical control elements in the hardware body are electrically connected with the matching control module. The matching control module includes a remote information input sub-module, a data storage sub-module, a triggering control sub-module, a medicine approval sub-module, and an intelligent analysis and decision sub-module. The remote information input sub-module obtains data related to patient diagnosis and treatment from the hospital end. The data storage sub-module collects and stores various data. The triggering control sub-module triggers the corresponding multimedia element based on the signals returned by the elements of the sensing unit. The intelligent analysis and decision sub-module formulates a personalized medication strategy based on data from the hospital end and the patient's physiological and life data.

2. The smart medication box based on patient discharge medication according to claim 1, wherein : The triggering control sub-module, Displays corresponding personalized information according to the differences in the person opening the medicine box. Displays personalized information according to the differences in the opened medicine cabin. Based on the returned information of the sensor, if the patient takes the wrong number or type of medicine, an alert is given.

3. The smart medication box based on patient discharge medication according to claim 1, wherein : The intelligent analysis and decision sub-module includes a patient personalized strategy self-construction unit, a multi-dimensional risk decision engine, and an AI question and answer unit, which form the optimal medication strategy through the fusion of the three.

4. The smart medication box based on patient discharge medication according to claim 3, wherein : The patient personalized model self-construction unit calls the information related to diagnosis and treatment, and formulates / regulates / corrects the patient's medication plan based on the medication rules of the medicine and the patient's own life habits and physiological adaptability.

5. The smart medication box based on patient discharge medication according to claim 3, wherein : The multi-dimensional risk decision engine identifies whether the patient is in a specific environment or scenario by monitoring the patient's physiological data, thereby determining whether the patient needs to adjust the medication strategy. The specific method is as follows: S1. Receive and fuse real-time physiological data, risk attributes of the medicine itself, patient personalized historical records, and related real-time and historical data; S2. Use weighted logistic regression or a lightweight neural network model to comprehensively calculate the above-mentioned multi-dimensional features to distinguish between physiological fluctuations and pathological abnormalities; S3. Connect the built-in clock and the weight sensor to compare the actual medicine taking time, the amount of medicine taken, and the preset medication plan to determine the impact of missed medication, wrong time medication, and similar behaviors on these physiological data; S4. Based on the above influence relationship, formulate a dynamic intervention strategy in different scenarios and levels.

6. The smart medication box based on patient discharge medication according to claim 3, wherein : The multi-dimensional risk decision engine analyzes the medical information of the patient during medication sent by the hospital end to determine whether the medication strategy needs to be adjusted. The specific method is as follows: S1. Receive the medical information from the hospital end; S2. Determine whether the patient's current medication conflicts with the ongoing or already performed or about to be performed recent medical event; S3. If there is a conflict, adjust the medication plan.

7. The smart medication box based on patient discharge medication according to claim 3, wherein : The AI question and answer unit provides patients with drug-related professional knowledge, captures abnormalities that cannot be monitored by other physiological monitoring devices during patient medication, and introduces them into the development of medication strategies.

8. The smart medication package based on patient discharge medication of claim 1, wherein : Also contains a drug approval submodule; The drug approval submodule counts the drug inventory during the drug placement stage and the drug inventory during the drug retrieval stage.

9. The smart medication package based on patient discharge medication of claim 1, wherein : Also contains a risk warning module; The risk warning module sends warning information to the patient / patient family member / hospital doctor end when an abnormal event occurs.

Images