A medication compliance monitoring and reminding platform for patients with chronic diseases in large internal medicine departments

Abstract

The application discloses a medication compliance monitoring and reminding platform for chronic disease patients in a large internal medicine department, and belongs to the technical field of medical intelligent monitoring. The platform comprises 10 modules in cooperation and linkage, adopts a distributed architecture for deployment, collects full-amount data related to patient medication through a multi-source data acquisition module, and transmits the data to a core operation and analysis module after processing by a special noise reduction formula. The core operation and analysis module realizes precise compliance evaluation, medication trend prediction and reminding strategy optimization based on multiple special mathematical formulas. An intelligent reminding module provides individualized and hierarchical reminding. A medical staff end management module realizes patient whole-process management and intervention scheme formulation. A data storage and encryption module ensures data security. A compliance intervention module realizes precise intervention and effect tracking. An abnormality early warning module realizes real-time early warning and linkage disposal. A cross-terminal synchronization module and a system self-checking and calibration module ensure stable operation of the system and data synchronization.

Description

Technical Field

[0001] This invention belongs to the field of medical intelligent monitoring technology, specifically involving a medication adherence monitoring and reminder platform for patients with chronic internal medicine diseases. Background Technology

[0002] As a comprehensive medical department, the Department of Internal Medicine treats a large number of patients with chronic diseases. Chronic diseases are characterized by long durations, long treatment cycles, multiple medications, and the need for long-term adherence to prescribed medications. Patient medication adherence directly affects treatment outcomes, disease control, and the risk of complications. Currently, medication adherence among chronic disease patients in the Department of Internal Medicine is generally low, mainly manifested as missed doses, incorrect doses, unauthorized discontinuation of medication, and improper dosage adjustments. This leads to relapses, poor treatment results, and even worsening of the condition, increasing medical costs and the burden on healthcare management.

[0003] Existing technologies include some medication reminder or adherence monitoring systems, but these systems generally have many shortcomings, failing to meet the medication management needs of patients with chronic internal medicine diseases and exhibiting significant technical limitations: First, the monitoring dimensions are singular, mostly only monitoring the patient's medication time, failing to combine multi-dimensional data such as physiological signs, medication environment, and medication awareness for comprehensive assessment, leading to inaccurate adherence assessments; second, the reminder strategies are fixed, using universal reminder methods and intensities, unable to be personalized according to the patient's adherence level, lifestyle habits, and physiological state, resulting in frequent false or missed reminders; third, the intervention measures lack specificity, employing uniform interventions. The proposed solutions suffer from several shortcomings: First, they fail to tailor interventions to individual patient conditions, including chronic disease type, duration, and factors influencing adherence, resulting in poor intervention outcomes. Second, data storage is insecure, lacking dedicated encryption mechanisms, posing a risk of patient privacy breaches, and exhibiting data redundancy and low query efficiency. Third, the system suffers from poor stability, insufficient computational precision, and cross-terminal data asynchrony, hindering long-term stable operation. Fourth, the lack of a dedicated adherence assessment system and computational model for internal medicine results in poor adaptability and an inability to accommodate the medication characteristics of various chronic diseases. Fifth, the inadequate early warning mechanism leads to untimely warnings, a high false alarm rate, and a lack of coordinated response capabilities, making it difficult to address medication and adherence abnormalities promptly.

[0004] Furthermore, most existing related systems rely on generalized hardware interfaces and software algorithms, resulting in severe technological homogenization and insufficient innovation. They are unable to achieve comprehensive and refined management of medication adherence for patients with chronic internal medicine diseases, and lack embedded proprietary mathematical formulas, resulting in simple operational logic, low assessment and prediction accuracy, and failing to meet the needs of clinical licensing and practical applications. Therefore, developing a medication adherence monitoring and reminder platform that can circumvent the shortcomings of existing technologies, adapt to the medication use scenarios of chronic internal medicine diseases, and achieve multi-dimensional monitoring, personalized reminders, precise intervention, and safe and stable operation has become an urgent technical problem to be solved. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing medication adherence monitoring systems, such as incomplete monitoring, inaccurate reminders, weak intervention targeting, data insecurity, system instability, poor adaptability, and insufficient computational precision. It provides a medication adherence monitoring and reminder platform for patients with chronic internal medicine diseases, specifically adapted to medication scenarios for various chronic diseases within the field of internal medicine. Through multi-module collaborative linkage, combined with a dedicated mathematical calculation model and encryption algorithm, it achieves full-process, multi-dimensional, high-precision monitoring of medication adherence, personalized reminders, precise intervention, and safe and stable operation, thereby improving patient medication adherence, enhancing treatment outcomes, and reducing the management burden on medical staff.

[0006] To achieve the above objectives, this invention provides the following technical solution: a medication adherence monitoring and reminder platform for patients with chronic internal medicine diseases, comprising a patient-side interaction module, a multi-source data acquisition module, a core computation and analysis module, an intelligent reminder module, a medical staff management module, a data storage encryption module, an adherence intervention module, an abnormality warning module, a cross-terminal synchronization module, and a system self-testing and calibration module. Each module is deployed in a distributed architecture, achieving bidirectional data interaction through a dedicated encryption protocol. Furthermore, each module operates independently yet collaboratively, without relying on the hardware interfaces and software of existing medication monitoring systems. This algorithm is specifically designed for medication scenarios in various chronic diseases within internal medicine (hypertension, diabetes, coronary heart disease, chronic bronchitis, sequelae of stroke, etc.). It enables full-process, multi-dimensional, and high-precision monitoring and personalized reminders of medication adherence. It also constructs a dedicated adherence assessment system and intervention logic. The specific architecture meets the following requirements: data transmission latency between modules ≤50ms, data interaction success rate ≥99.92%, continuous system stability ≥99.87%, and all data calculations are based on a dedicated mathematical model, avoiding the single-dimensional monitoring and generalized reminder logic of existing technologies.

[0007] Furthermore, the multi-source data acquisition module is used to collect comprehensive medication-related data from patients with chronic internal medicine diseases, avoiding the single data acquisition mode in existing technologies. Specifically, the collected data includes patient basic information, medication orders, real-time medication behavior data, physiological sign correlation data, medication environment data, and medication feedback data. Patient basic information includes age, gender, disease duration, complication type, drug allergy history, liver and kidney function indicators, education level, and medication awareness. Medication orders include drug type, specifications, single dose, frequency, medication time window, treatment course, discontinuation criteria, and dosage adjustment rules. Real-time medication behavior data... Data is collected using non-invasive sensing technology, including actual medication time, actual dosage, number of missed doses, number of incorrect doses, duration of unauthorized discontinuation, and dosage adjustment behavior; physiological indicators directly related to medication efficacy, such as blood pressure, blood glucose, heart rate, blood oxygen saturation, and respiratory rate, are collected every 30 minutes, with the frequency increased to every 5 minutes under abnormal conditions; environmental data includes ambient temperature and humidity, and light intensity (to avoid misjudgments of medication adherence due to improper drug storage); medication feedback data includes post-medication discomfort symptoms, evaluation of medication convenience, and feedback on factors affecting adherence; a noise reduction algorithm is used during data collection, and the noise reduction formula is as follows: ,in These are the data values ​​after noise reduction. These are the original collected data values. The sample size for a single data collection group ( ≥30), This is the average of the original data. The noise reduction coefficient (range 0.02-0.08, dynamically adjusted according to data type) ensures that the accuracy of the collected data is ≥99.7%, and the collection method does not affect the patient's normal life, avoiding the problems of existing invasive collection and data omission and miscollection.

[0008] Furthermore, the core computation and analysis module, as the core processing unit of the platform, avoids the generalized computational logic and single evaluation model in existing technologies. It adopts a multi-dimensional fusion computational algorithm, based on the full amount of data collected by the multi-source data acquisition module, to achieve accurate quantitative assessment of patient medication adherence, prediction of medication behavior trends, and dynamic optimization of reminder strategies. Specifically, it includes three sub-units: a adherence quantitative assessment unit, a medication trend prediction unit, and a reminder strategy optimization unit. The adherence quantitative assessment unit constructs a specific adherence assessment index system for chronic internal medicine diseases. The indicators include six core indicators: medication on-time rate, dosage accuracy rate, treatment course completion rate, missed dose frequency, accidental dose frequency, and duration of unauthorized discontinuation of medication. The weights of each indicator are determined using the analytic hierarchy process (AHP), and the weight calculation formula is as follows: ; in For the first Weight of each indicator ( =1,2,...,6), For the first A senior internal medicine expert on the first Scoring of each indicator ( ≥15 (score range 1-10), and meets the following conditions The final compliance score calculation formula is as follows: ,in The compliance score is calculated as follows (range 0-100 points). For the first The actual score of each indicator (range 0-100 points); based on the comprehensive score, compliance is divided into 5 levels: high compliance ( ≥90 points), high compliance (80 points ≤ <90 points), moderate compliance (65 points ≤ <80 points), low compliance (50 points ≤ <65 points), low compliance ( <50 points), each level corresponds to a specific intervention threshold and reminder intensity, avoiding the problems of overly coarse compliance grading and inaccurate assessment in existing technologies.

[0009] Furthermore, the medication trend prediction unit employs an improved time-series prediction algorithm to overcome the limitations of existing single-trend prediction models. Based on patients' historical medication behavior data (containing at least 3 months of complete data), it predicts the trend of changes in patients' medication adherence over the next 1-4 weeks. The prediction formula is as follows: ,in For the first Zhou's compliance prediction score For the first Zhou's actual compliance score This is a trend adjustment factor (range 0.05-0.12, dynamically adjusted according to the duration of the disease; the longer the disease duration, the higher the trend adjustment factor). (The smaller the value) To backtrack the number of weeks for historical data ( ≥12), For the first The difference in weekly physiological signs (calculated by comparing current physiological sign data with the average of previous periods). The physiological sign influence coefficient (range 0.10-0.18, dynamically adjusted according to the type of chronic disease, including patients with diabetes and hypertension). (Value ≥ 0.15); prediction error controlled within ±3.5 points, prediction accuracy ≥ 92.3%, capable of identifying the risk of declining adherence in advance, providing data support for the implementation of subsequent intervention measures; the reminder strategy optimization unit dynamically optimizes the reminder parameters based on adherence scores, medication trend prediction results, and patient medication feedback data, with the optimization formula as follows: The optimized reminder lead time (in minutes). The standard reminder time is set in advance (default 30 minutes, which can be adjusted by medical staff). The current overall compliance score, To predict compliance scores, To optimize the coefficient (ranging from 0.03 to 0.07), and ensure that the reminder strategy is highly compatible with patients' medication habits and compliance levels, the problem of fixed reminder strategies and inability to be dynamically adjusted in existing technologies is avoided.

[0010] Furthermore, the intelligent reminder module avoids the problems of single-form reminders, fixed reminder intensity, and high false alarm rates in existing technologies. It adopts a multi-mode integrated reminder approach, combining patient compliance level, medication scenario, physiological signs, and medication feedback to achieve personalized and hierarchical reminders. Specifically, it includes three sub-units: a reminder form adaptation unit, a reminder intensity adjustment unit, and a false alarm avoidance unit. The reminder forms include voice reminders, text reminders, vibration reminders, light reminders, and reminders linked to the medical staff's device. Patients or medical staff can choose according to the patient's lifestyle (e.g., whether they have hearing impairment or are asleep), and it supports simultaneous reminders of multiple forms. The reminder intensity adjustment unit dynamically adjusts the reminder intensity based on the compliance level, with the intensity grading formula as follows: ,in The current alert level (range 1-5). The lowest alert level (Level 1). This is the highest alert level (Level 5). The current overall adherence score is used. The lower the adherence, the higher the reminder intensity, and the reminder interval gradually shortens (30-minute reminder interval for high adherence, 5-minute reminder interval for low adherence, until the patient confirms medication use). The false reminder avoidance unit determines whether the patient has completed medication use by identifying the patient's medication behavior data and physiological sign data, avoiding duplicate and false reminders. The determination formula is: ; in This indicates that the medication has been taken and no further reminder is needed. This indicates that the medication has not been completed and a reminder is needed. This refers to the actual time of medication administration. For standard medication time, This is the actual dosage. For standard medication dosage, Physiological data at the time of medication administration. , Within the normal range of physiological signs, the false alarm rate should be controlled at ≤0.8%.

[0011] Furthermore, the aforementioned medical staff management module avoids the problems of existing medical staff terminals having limited functionality and being unable to achieve precise intervention and batch management. It is specifically adapted to the work scenarios of internal medicine medical staff, enabling full-process management, precise intervention, and batch operations for medication adherence of patients under their jurisdiction. Specifically, it includes five sub-units: patient management, adherence viewing, intervention plan formulation, batch operation, and data statistical analysis. The patient management unit supports medical staff in classifying and managing patients according to chronic disease type, adherence level, disease duration, and ward. It allows adding, editing, and deleting patient information, viewing complete patient medication records (including basic information, medical orders, medication behavior data, physiological signs data, adherence scores, intervention records, etc.), and supports rapid retrieval and fuzzy search of patient information with a retrieval response time of ≤1 second. Adherence... The viewing unit allows healthcare professionals to view a single patient's real-time adherence score, historical adherence change curve, specific scores for various assessment indicators, and analysis of the reasons for adherence changes (automatically generated based on medication behavior data and physiological sign data). The intervention plan development unit allows healthcare professionals to develop personalized adherence intervention plans based on patients' adherence levels, medication trends, physiological signs, and complications. These plans include medication guidance, psychological counseling suggestions, dosage adjustment reminders, and follow-up plans, and the intervention plans can be automatically synchronized to the patient's end for real-time tracking of intervention effects. The batch operation unit supports healthcare professionals in simultaneously reminding multiple patients, batch distributing intervention plans, and batch developing follow-up plans, improving healthcare management efficiency. The data statistical analysis unit automatically calculates the overall adherence of patients under their management, using statistical formulas including: the overall adherence mean formula. (in The overall compliance mean. For the first Patient compliance scores, Formula for total number of patients under management and percentage of each compliance level. (in For the first The percentage of each compliance level For the first The number of patients at each compliance level =1,2,...,5), the statistical results are displayed in the form of charts and graphs, and statistical reports can be exported, providing data support for the management of medication for chronic diseases in internal medicine.

[0012] Furthermore, the data storage encryption module avoids the problems of insecure data storage, data redundancy, low query efficiency, and high privacy leakage risks in existing technologies. It adopts a layered encrypted storage architecture combined with a dedicated encryption algorithm to achieve secure storage, efficient querying, and privacy protection of all patient medication-related data. Specifically, it includes four sub-units: a data layered storage unit, a data encryption unit, a data backup unit, and a data cleaning unit. The data layered storage unit divides the data into three levels: core data (patient basic information, medical order information, adherence scores), general data (medication behavior data, physiological sign data), and temporary data (medication environment data, temporary medication feedback records). Each level uses a different storage method: core data uses a distributed database, general data uses a relational database, and temporary data uses a cache. The storage capacity is dynamically expandable, supporting a single patient data storage duration of ≥5 years and a data query response time of ≤0.5s. The data encryption unit uses a dedicated symmetric encryption algorithm to encrypt the core data throughout the entire process. The encryption formula is: ,in For encrypted data, For the original core data, Use a dedicated encryption key (automatically generated by the system, updated every 30 days, key length ≥ 256 bits). The encryption offset (values ​​ranging from 100 to 200, randomly generated) is used to decrypt the following formula: To ensure core data is not leaked or tampered with during storage and transmission, the data backup unit employs a combination of real-time and scheduled backups. Real-time backups are performed every minute, while scheduled backups are performed daily at 2:00 AM. Backup data is stored on multiple off-site servers, achieving a backup success rate of ≥99.99%. It supports data recovery from accidental deletion (recovery time ≤10 minutes). The data cleanup unit automatically cleans up expired temporary data (cleanup cycle of 7 days) and compresses redundant data with a compression ratio of ≥80%, ensuring the efficient operation of the storage system. Simultaneously, it strictly adheres to medical data privacy protection regulations, authorizing only medical staff and the patients themselves to view patient data, thus mitigating the risk of privacy leaks.

[0013] Furthermore, the adherence intervention module avoids the problems of single intervention measures, weak targeting, and poor intervention effects in existing technologies. Based on the adherence score, medication trend prediction results, and intervention plans formulated by medical staff obtained from the core computing and analysis module, it realizes personalized, full-process intervention for patients' medication adherence. Specifically, it includes three sub-units: an intervention measure matching unit, an intervention effect tracking unit, and an intervention plan adjustment unit. The intervention measure matching unit matches corresponding intervention measures according to the patient's adherence level, chronic disease type, disease duration, medication awareness level, and factors affecting adherence. Patients with high adherence only receive medication reminder reinforcement services, while patients with low or no adherence receive comprehensive intervention measures such as medication guidance, psychological counseling, family interaction, and follow-up reminders. Moreover, the intervention measures can be dynamically adjusted based on patient feedback. The intervention effect tracking unit tracks the implementation effect of the intervention measures in real time, using an intervention effect evaluation formula: ,in The intervention effect rate (range: -100% to +100%). For adherence scores after intervention, For compliance scores before intervention, ≥10% indicates effective intervention, E<0% indicates ineffective intervention, requiring adjustment of the intervention plan; the intervention plan adjustment unit automatically or manually adjusts the intervention plan based on the intervention effect evaluation results, patient medication feedback, and changes in medication trends. The adjustment principle is: if ≥20%, maintain the current intervention plan; if 10% ≤ If <20%, optimize the frequency of intervention implementation; if If the rate is less than 10%, change the type of intervention to ensure that the intervention effect continues to improve, and ultimately achieve the goal of increasing patient compliance by an average of ≥15%.

[0014] Furthermore, the aforementioned anomaly warning module avoids the problems of untimely warnings, fixed warning thresholds, high false alarm rates, and lack of coordinated response in existing technologies. Based on medication behavior data and physiological sign data collected by the multi-source data acquisition module and the calculation results of the core calculation and analysis module, it achieves real-time warning and coordinated response for medication abnormalities, compliance abnormalities, and physiological sign abnormalities. Specifically, it includes three sub-units: an anomaly identification unit, a warning threshold dynamic adjustment unit, and a warning coordinated response unit. The anomaly types include: missed doses ≥3 consecutive times, dosage deviation ≥10%, unauthorized discontinuation of medication ≥24 hours, a sudden drop in compliance score ≥15 points (within 24 hours), and physiological sign data exceeding the normal range for ≥10 minutes. The anomaly identification unit uses an anomaly identification algorithm, with the following formula: ; in This indicates an anomaly and requires triggering an alert. This indicates no abnormalities and no warning is needed. This represents the number of consecutive missed doses. For the duration of unauthorized discontinuation of medication, For the current compliance score, The adherence score is based on data from 24 hours prior; the warning threshold is dynamically adjusted based on the patient's disease duration, complication type, and medication history data. The adjustment formula is as follows: ,in The optimized warning threshold, The standard warning threshold is... The warning threshold is set at the duration of the illness (in years). The longer the illness duration, the more lenient the warning threshold, to avoid excessive warnings. The warning linkage and response unit immediately sends a warning reminder (at the highest intensity) to the patient's end after the warning is triggered, and at the same time sends warning information (including abnormality type, abnormal data, and patient information) to the corresponding medical staff end. If no action is taken within 30 minutes after the warning is issued, the system will automatically contact the patient's family (with pre-reserved contact information) and generate a warning response record, recording the warning time, abnormality type, response method, and response effect. The false alarm rate is controlled at ≤1.2%, and the warning response time is ≤10s.

[0015] Furthermore, the cross-terminal synchronization module and the system self-test calibration module work together to avoid the problems of cross-terminal data asynchrony, system instability, and decreased computational accuracy in existing technologies. The cross-terminal synchronization module supports real-time data synchronization between patient terminals (mobile phones, tablets, smartwatches), medical staff terminals (computers, mobile phones), and the server, using a dedicated synchronization protocol. The synchronization formula is as follows: ,in For the synchronized data, For data from the data source terminal, For the target terminal's data, The synchronization weight is set to 0.8-0.9, with higher priority data source terminals. (The larger the value), the synchronization delay is ≤100ms, and the data synchronization accuracy is ≥99.95%, ensuring that the data viewed by patients and medical staff is consistent; the system self-test calibration module adopts a combination of periodic self-testing and real-time calibration, with a self-testing cycle of once every 2 hours. The self-testing content includes the operating status of each module, data transmission status, calculation accuracy, and storage system status. The formula for judging whether the self-test is qualified is: ; in Self=0 indicates that the self-test is passed; Self=0 indicates that the self-test is failed. This indicates the module's operating status (1 indicates normal operation, 0 indicates an error). For data transmission delay, This is due to computational precision error. To store remaining capacity; if the self-test fails, an immediate system alarm will be triggered, and automatic calibration will be performed simultaneously. The calibration formula is: Where Cal is the calibrated operational parameter, These are standard operational parameters. To minimize computational accuracy errors, the calibration success rate is ≥99.8%. If calibration fails, the system automatically switches to the backup system to ensure continuous and stable operation of the platform and that the computational accuracy always meets the requirements.

[0016] This invention provides a medication adherence monitoring and reminder platform for patients with chronic internal medicine diseases, which has the following beneficial effects: (1) Avoid the single-dimensional monitoring defects in existing technologies. Use a multi-source data acquisition module to collect full data such as patient basic information, medication orders, medication behavior, physiological signs, medication environment and medication feedback. Combined with a dedicated noise reduction formula, ensure the accuracy and comprehensiveness of the collected data, provide reliable data support for compliance assessment, and adapt to medication scenarios for various chronic diseases in internal medicine.

[0017] (2) By avoiding the generalized calculation logic in existing technologies, a specific compliance assessment system and core calculation model for chronic internal medicine diseases are constructed, and multiple specific mathematical formulas (compliance scoring formula, trend prediction formula, reminder intensity adjustment formula, etc.) are embedded to achieve accurate quantitative assessment of compliance and accurate prediction of medication trends. The prediction accuracy rate is ≥92.3%, and the compliance assessment error is ≤3.5 points, which is significantly better than existing technologies.

[0018] (3) To avoid the defects of fixed reminders in existing technologies, a multi-mode integrated reminder method is adopted. The intensity and interval of reminders are dynamically adjusted in combination with the patient's compliance level. Through the false reminder avoidance formula, the false reminder rate is controlled at ≤0.8%, so as to realize personalized and precise reminders, improve the patient's medication experience, and reduce missed doses and false doses.

[0019] (4) Avoiding the shortcomings of the single management function of the medical and nursing terminal in the existing technology, the medical and nursing terminal management module supports patient classification management, compliance viewing, personalized intervention plan formulation, batch operation and data statistical analysis. Combined with exclusive statistical formulas, it provides data support for the management of medication for chronic diseases in internal medicine and improves the efficiency of medical and nursing management.

[0020] (5) Avoid the data security risks in existing technologies, adopt a layered encrypted storage architecture and exclusive encryption and decryption formulas, and combine real-time backup and regular backup mechanisms to ensure the safe storage and privacy protection of patient data, strictly follow the relevant regulations on medical data privacy protection, and the risk of data leakage is close to 0.

[0021] (6) Avoid the shortcomings of existing technologies in terms of weak targeting and poor intervention effect, construct a personalized intervention system, combine with a special intervention effect evaluation formula, track the intervention effect in real time and dynamically adjust the intervention plan to ensure that the patient compliance is improved by an average of ≥15% and improve the treatment effect of chronic diseases.

[0022] (7) To avoid the shortcomings of untimely early warning and high false alarm rate in the existing technology, an anomaly identification formula and a dynamic adjustment formula for early warning threshold are adopted to realize real-time identification and accurate early warning of abnormal situations. The false alarm rate is controlled at ≤1.2%, the early warning response time is ≤10s, and it has the ability to handle various abnormal situations in a timely manner.

[0023] (8) By avoiding the defects of cross-terminal data asynchrony, system instability and decreased calculation accuracy in the existing technology, the cross-terminal synchronization module and the system self-test calibration module, combined with the exclusive synchronization formula and calibration formula, ensure that the cross-terminal data synchronization accuracy is ≥99.95%, the system continuous operation stability is ≥99.87%, and the calculation accuracy always meets the clinical needs.

[0024] (9) It does not rely on the hardware interface and software algorithm of the existing medication monitoring system. All modules, operation logic and mathematical formulas are specially designed, with significant novelty and creativity, and are suitable for medication scenarios of various chronic diseases in internal medicine. Attached Figure Description

[0025] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings in the following description are merely exemplary, and those skilled in the art can derive other embodiments based on the provided drawings without creative effort.

[0026] Figure 1 This is a flowchart of the multi-source data acquisition process of the present invention; Figure 2 This is a flowchart illustrating the core computational analysis process of the present invention. Figure 3 This is a flowchart of the intelligent reminder process of the present invention; Figure 4 This is a flowchart of the medical staff management process of this invention. Detailed Implementation

[0027] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses consistent with some aspects of this disclosure as detailed in the appended claims.

[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0029] How to use:

[0030] 1. Platform Startup and Initialization: After the platform is started, the system automatically starts the 10 modules described in claim 1. Each module completes bidirectional interactive initialization through a dedicated encryption protocol to ensure that data transmission delay, interaction success rate, and system stability meet the standards specified in claim 1. After initialization is completed, it enters standby mode to wait for subsequent operations.

[0031] 2. Multi-source data acquisition module operation: Activate the multi-source data acquisition module described in claim 2. Medical staff can use this module to input basic patient information and medication order data, and patients can cooperate to complete the input of medication feedback data. The module automatically collects real-time medication behavior data through non-invasive sensing technology, and collects physiological sign-related data and medication environment data at a specified frequency. During the acquisition process, the module automatically runs a dedicated noise reduction formula to complete data noise reduction. After ensuring the accuracy of the collected data, it is transmitted to the core computing and analysis module.

[0032] 3. Operation of the core computing and analysis module: After receiving data transmitted by the multi-source data acquisition module, the core computing and analysis module described in claim 3 automatically starts the operation of three sub-units: the compliance quantification assessment unit completes the patient's medication compliance score and level classification through a dedicated weight calculation formula and a comprehensive compliance scoring formula; the medication trend prediction unit completes the compliance trend prediction for the next 1-4 weeks through an improved time-series prediction formula; and the reminder strategy optimization unit generates personalized reminder parameters through a reminder strategy optimization formula. All calculation results are synchronized to the intelligent reminder module and the medical staff management module.

[0033] 4. Use of the intelligent reminder module: After receiving the optimized parameters from the core computing and analysis module, the intelligent reminder module described in claim 5 automatically starts three sub-units: the reminder form adaptation unit determines the reminder form and completes the adaptation according to the patient's preset preferences; the reminder intensity adjustment unit adjusts the reminder intensity and interval according to the patient's compliance level through the reminder intensity grading formula; and the false reminder avoidance unit judges the patient's medication status through the false reminder avoidance formula, avoids false reminders and repeated reminders, and automatically sends a medication reminder of the corresponding form and intensity after the preset medication time, until the patient confirms medication.

[0034] 5. Operation of the Medical Staff Management Module: Medical staff log in to the medical staff management module described in claim 6, and complete patient classification management, information retrieval, and file viewing through the patient management unit; view real-time patient compliance scores, historical change curves, and indicator scores through the compliance viewing unit; formulate personalized intervention plans based on the calculation results of the core calculation and analysis module through the intervention plan formulation unit and synchronize them to the patient terminal; complete batch reminders and intervention plan distribution through the batch operation unit; and generate an overall patient compliance statistical report by running exclusive statistical formulas through the data statistical analysis unit.

[0035] 6. Data storage encryption module operation: The data storage encryption module described in claim 7 automatically receives data transmitted from each module and classifies and stores it according to the hierarchical standard of core data, ordinary data, and temporary data; it encrypts the core data using a dedicated encryption formula, automatically performs real-time backup and periodic backup operations, cleans up expired temporary data and compresses redundant data according to the prescribed cycle, ensuring data security and efficient operation of the storage system, and authorizing medical staff and patients to query the corresponding data.

[0036] 7. Compliance Intervention Module Operation: After receiving the calculation results from the core calculation and analysis module and the intervention plan from the medical staff, the compliance intervention module described in claim 8 matches the corresponding intervention measures through the intervention measure matching unit and executes them; through the intervention effect tracking unit, it runs the intervention effect evaluation formula to track the intervention effect; based on the evaluation results, through the intervention plan adjustment unit, it dynamically adjusts the intervention plan according to the corresponding principle to continuously improve patients' medication compliance.

[0037] 8. Operation of the abnormality early warning module: The abnormality early warning module described in claim 9 receives data from the multi-source data acquisition module and the core computing and analysis module in real time. It runs the abnormality identification formula through the abnormality identification unit to identify various medication, compliance, and physiological abnormalities. It dynamically optimizes the early warning threshold through the dynamic adjustment formula. After an abnormality is identified, it sends early warning information to the patient and medical staff through the early warning linkage and handling unit. If no action is taken within the specified time, it automatically contacts the patient's family and records the early warning handling situation.

[0038] 9. Cross-terminal synchronization and system self-calibration: The cross-terminal synchronization module described in claim 10 automatically runs a dedicated synchronization formula to achieve real-time data synchronization between the patient, medical staff, and server terminals; the system self-calibration module performs self-checks at prescribed intervals, determines the system's operating status using a self-check pass / fail judgment formula, and automatically completes calibration using a calibration formula if the self-check fails. If calibration fails, the system switches to a backup system to ensure continuous and stable platform operation.

[0039] Example 1: Platform Application for Hypertensive Patients in Internal Medicine

[0040] This embodiment is adapted to all 10 modules of the platform described in claim 1, and is specifically applied to the scenario of medication adherence monitoring and reminders for hypertension patients in internal medicine. It follows the platform usage method throughout the process and strictly corresponds to the functions of each module and the claim numbers.

[0041] After the platform starts and initializes, each module completes the interaction preparation according to the collaborative logic specified in claim 1 and enters the standby state. The multi-source data acquisition module described in claim 2 is started. Medical staff enter the basic information and medication orders of hypertensive patients through this module. Patients complete the medication feedback data entry through the patient-side interaction module based on their own medication experience. This module automatically collects the patient's real-time medication behavior data through non-invasive sensing technology, collects physiological sign correlation data related to the effect of hypertension medication at a preset frequency, and collects data related to the patient's medication environment. During the collection process, a dedicated noise reduction formula is automatically run to denoise various types of raw collected data to ensure the integrity and accuracy of the collected data. After processing, all data is synchronously transmitted to the core computing and analysis module.

[0042] After receiving data, the core computation and analysis module described in claim 3 automatically starts the collaborative operation of its three sub-units. The compliance quantification assessment unit, based on the assessment index system specified in claim 3, determines the weight of each core index using a dedicated weighting formula, then completes the comprehensive medication compliance score for the hypertensive patient using a comprehensive compliance scoring formula, and determines the patient's compliance level according to the set grading standards. The medication trend prediction unit, using an improved time-series prediction formula and combining the patient's historical medication data, predicts the trend of changes in the patient's medication compliance over a future period. The reminder strategy optimization unit, combining the current compliance score, trend prediction results, and patient medication feedback data, generates personalized reminder parameters tailored to the patient using a dedicated reminder strategy optimization formula. All computation results are synchronously transmitted to the intelligent reminder module and the medical staff management module.

[0043] After receiving the optimized reminder parameters, the intelligent reminder module described in claim 5 automatically starts the operation of three sub-units. The reminder format adaptation unit selects a reminder format that suits the hypertensive patient's lifestyle based on their pre-set preferences and completes the adaptation. The reminder intensity adjustment unit dynamically adjusts the reminder intensity and interval using a reminder intensity grading formula, combined with the patient's current compliance level. The false reminder avoidance unit uses a false reminder avoidance formula to judge the patient's actual medication status in real time, effectively avoiding repeated reminders and false reminders. When the preset medication time is reached, it automatically sends a medication reminder of the corresponding format and intensity until the patient confirms completion of the medication operation.

[0044] Medical staff can log in to the medical staff management module described in claim 6 through a terminal device. Through the patient management unit, they can classify and manage hypertensive patients, quickly retrieve patient information, and view complete medication records. Through the adherence viewing unit, they can view the patient's comprehensive medication adherence score, historical adherence change curve, and specific scores of various assessment indicators in real time, clearly understanding the patient's medication adherence status. Based on the calculation results transmitted by the core calculation and analysis module, and combined with the patient's condition characteristics, the intervention plan formulation unit can formulate targeted adherence intervention plans and synchronize them to the patient's terminal. If managing multiple hypertensive patients, the batch operation unit can complete batch medication reminders and batch distribution of intervention plans, improving management efficiency. Through the data statistical analysis unit, dedicated statistical formulas can be run to statistically analyze the overall medication adherence of the managed hypertensive patients, generating statistical reports to support the management of hypertension medication in internal medicine.

[0045] The data storage encryption module described in claim 7 automatically receives all data transmitted from each module and classifies and stores the data according to a hierarchical standard of core data, ordinary data, and temporary data. Core data is stored using a distributed database, ordinary data is stored using a relational database, and temporary data is stored using a cache. The core data is encrypted throughout the process using a dedicated encryption formula. At the same time, real-time backup and periodic backup operations are automatically performed. Expired temporary data is cleaned up according to a preset cycle, and redundant data is compressed to ensure the secure storage and privacy protection of patient data. Only medical staff and the patient are authorized to view the corresponding data.

[0046] After receiving the calculation results from the core calculation and analysis module and the intervention plan formulated by the medical staff, the compliance intervention module described in claim 8, through the intervention measure matching unit, combines the patient's compliance level, disease characteristics, and compliance influencing factors to match the corresponding intervention measures and implement them; through the intervention effect tracking unit, it runs the intervention effect evaluation formula to track the implementation effect of the intervention measures in real time; based on the intervention effect evaluation results, patient medication feedback, and changes in medication compliance trends, through the intervention plan adjustment unit, it dynamically adjusts the intervention plan according to preset principles to continuously improve the patient's medication compliance.

[0047] The abnormality warning module described in claim 9 receives medication behavior data, physiological sign correlation data, and calculation results from the core calculation and analysis module collected by the multi-source data acquisition module in real time. It then uses an abnormality identification unit to run an abnormality identification formula to identify abnormal medication use, compliance, and physiological signs in patients in real time. The warning threshold is dynamically optimized by the warning threshold adjustment unit, taking into account the patient's disease duration and condition characteristics. When an abnormality is detected, the warning linkage and handling unit immediately sends the strongest warning alert to the patient and simultaneously sends a warning message to the corresponding medical staff. If no action is taken within a preset time after the warning is issued, the module automatically contacts the patient's family and records the warning time, abnormality type, handling method, and handling effect in detail.

[0048] The cross-terminal synchronization module described in claim 10 automatically runs a dedicated synchronization formula to synchronize all data between the patient terminal, medical staff terminal, and server in real time, ensuring that the data viewed by patients and medical staff remains consistent. The system self-test calibration module performs self-test operations according to a preset cycle, and judges the operating status, data transmission status, calculation accuracy, and storage system status of each module through a self-test pass / fail judgment formula. If the self-test fails, the system is immediately calibrated automatically through a calibration formula. If the calibration fails, the system is automatically switched to a backup system to ensure the continuous and stable operation of the platform and to ensure the orderly implementation of medication adherence monitoring and reminders for hypertension patients throughout the entire process.

[0049] Example 2: Platform Application for Diabetic Patients in Internal Medicine

[0050] This embodiment is also based on all 10 modules of the platform described in claim 1, and is specifically applied to the scenario of medication adherence monitoring and reminders for diabetic patients in internal medicine. It strictly follows the platform usage method, corresponds to the functions of each module and the claim reference numbers, avoids existing technologies, and does not contain any specific data.

[0051] After the platform starts, it automatically completes the initialization operation. The 10 modules described in claim 1 complete bidirectional interaction through a dedicated encryption protocol to ensure that the operating parameters of each module meet the standards specified in claim 1. After initialization, it enters a standby state, waiting for subsequent operation instructions. The multi-source data acquisition module described in claim 2 is then activated. Medical staff use this module to enter the basic information and medication orders of diabetic patients, clarifying the patients' medication-related requirements. Patients use the patient-side interaction module to periodically provide feedback on their feelings and related conditions after taking medication, completing the medication feedback data entry. This module automatically collects the patient's real-time medication behavior data through non-invasive sensing technology, capturing the patient's medication actions and related behaviors, collecting physiological sign correlation data related to diabetes medication at a preset frequency, and simultaneously collecting relevant data on the patient's medication environment. During the collection process, a dedicated noise reduction formula is automatically run to denoise the collected raw data, remove invalid interference data, and ensure the accuracy of the data before transmitting all processed datasets to the core computing and analysis module.

[0052] After receiving the dataset transmitted by the multi-source data acquisition module, the core operation and analysis module described in claim 3 initiates collaborative operations by its three sub-units. The adherence quantification assessment unit determines the weight of each indicator using a proprietary weighting formula based on the six core assessment indicators specified in claim 3. Combining this with the actual scores of each indicator, it calculates the comprehensive medication adherence score for the diabetic patient using a comprehensive adherence scoring formula and assigns corresponding adherence levels according to the scoring criteria. The medication trend prediction unit, based on the patient's historical medication behavior data and current medication status, predicts the future trend of medication adherence using an improved time-series prediction formula, identifying the risk of declining adherence in advance. The reminder strategy optimization unit, combining the patient's current adherence score, trend prediction results, and medication feedback data, optimizes parameters such as reminder lead time using a proprietary reminder strategy optimization formula, generating a personalized reminder strategy tailored to the diabetic patient. All calculation results are simultaneously sent to the intelligent reminder module and the healthcare management module.

[0053] After receiving the optimization parameters sent by the core computation and analysis module, the intelligent reminder module described in claim 5 starts three sub-units to operate. The reminder format adaptation unit selects a suitable reminder format based on the diabetic patient's lifestyle and individual circumstances, supporting combinations of multiple reminder formats to meet the patient's personalized needs. The reminder intensity adjustment unit dynamically adjusts the reminder intensity and interval based on the patient's compliance level using a reminder intensity grading formula; the lower the compliance, the higher the reminder intensity and the shorter the interval. The false reminder avoidance unit accurately determines whether the patient has completed medication administration using a false reminder avoidance formula, combined with the patient's real-time medication behavior data and physiological sign data, effectively avoiding false and repeated reminders. When the preset medication time is reached, it automatically triggers a medication reminder of the corresponding format and intensity until the patient confirms medication administration.

[0054] Medical staff can log in to the medical staff management module described in claim 6, and classify and manage diabetic patients according to chronic disease type through the patient management unit. They can quickly edit and query relevant patient information, and view the patient's complete medication record, including basic information, medical order information, medication behavior data, and physiological sign data. Through the adherence viewing unit, they can view the patient's comprehensive medication adherence score and historical change curve in real time to understand the changes in the patient's medication adherence and the scores of various assessment indicators. Based on the calculation results of the core calculation and analysis module, combined with the actual condition of the patient, personalized adherence intervention plans can be formulated through the intervention plan formulation unit, clarifying the intervention measures and implementation requirements, and synchronized to the patient's end so that the patient can clearly understand the intervention focus. Through the batch operation unit, multiple diabetic patients can be managed in batches, improving the efficiency of medical staff. Through the data statistical analysis unit, dedicated statistical formulas can be run to statistically analyze the overall adherence of the diabetic patients under their management, generating intuitive statistical charts and reports, providing data support for the management of diabetes medication in internal medicine.

[0055] The data storage encryption module described in claim 7 receives data transmitted from each module in real time, classifies and stores the data according to hierarchical storage standards to ensure that the data storage of various types is standardized and efficient; it encrypts core data using a dedicated encryption formula to protect patient privacy and data security, and the encryption key is automatically updated according to a preset period; it also performs real-time backup and periodic backup operations, stores backup data on a remote server to prevent data loss, cleans up expired temporary data according to a preset period, and compresses redundant data to ensure the efficient operation of the storage system and strictly comply with relevant regulations on medical data privacy protection.

[0056] After receiving the calculation results from the core calculation and analysis module and the intervention plan from the medical staff, the compliance intervention module described in claim 8, through the intervention measure matching unit, combines the patient's compliance level, disease characteristics, and medication awareness level to match and execute corresponding intervention measures; through the intervention effect tracking unit, it runs the intervention effect evaluation formula to periodically evaluate the implementation effect of the intervention measures and determine whether the intervention is effective; based on the evaluation results, patient medication feedback, and changes in medication compliance trends, through the intervention plan adjustment unit, it dynamically adjusts the intervention plan, optimizes the implementation method and frequency of the intervention measures, ensures continuous improvement of the intervention effect, and helps patients improve medication compliance.

[0057] The abnormality early warning module described in claim 9 monitors the data transmitted by the multi-source data acquisition module and the core computing and analysis module in real time. Through the abnormality identification unit, it runs the abnormality identification formula to identify medication abnormalities such as missed doses, accidental doses, and unauthorized discontinuation of medication, as well as sudden drops in compliance and abnormal physiological signs. Through the early warning threshold dynamic adjustment unit, it dynamically adjusts the early warning threshold in combination with the patient's disease duration and complications to avoid excessive or untimely early warnings. When an abnormality is identified, it immediately sends an early warning reminder to the patient and sends detailed early warning information to the medical staff through the early warning linkage and handling unit. If no action is taken within a preset time after the early warning is issued, it automatically contacts the patient's family and records the entire early warning handling process to form a complete early warning handling file.

[0058] The cross-terminal synchronization module described in claim 10 automatically runs a dedicated synchronization formula to achieve real-time data synchronization between the patient end, the medical staff end, and the server, ensuring that patients and medical staff can always view the latest data and avoiding management errors caused by data asynchrony. The system self-test calibration module performs self-tests on the entire platform according to a preset cycle. It comprehensively tests the operating status, data transmission latency, calculation accuracy, and remaining storage capacity of each module through a self-test pass / fail judgment formula. If a self-test failure is found, the system is automatically calibrated immediately through a calibration formula. If the calibration fails, the system is automatically switched to a backup system to ensure the continuous and stable operation of the platform and provide reliable support for the monitoring and reminder of medication adherence for diabetic patients.

[0059] Example 3: Platform Application for Mixed Management of Multiple Chronic Diseases in Internal Medicine

[0060] This embodiment is based on all 10 modules of the platform described in claim 1 and is applied to the scenario of mixed management of patients with multiple chronic diseases in internal medicine. It can simultaneously adapt to the medication adherence monitoring and reminder needs of patients with multiple chronic diseases in internal medicine such as hypertension, diabetes, and coronary heart disease. It strictly follows the platform usage method, corresponds to the functions of each module and the claim reference numbers, avoids existing technologies, and does not contain any specific data.

[0061] After the platform is started, it automatically starts the 10 modules described in claim 1. Each module adopts a distributed architecture and operates collaboratively. It completes bidirectional data interaction through a dedicated encryption protocol and completes the initialization operation to ensure that data transmission delay, interaction success rate, system stability and other aspects meet the standards specified in claim 1. After initialization, it enters standby mode and can receive relevant operations from multiple patients with different types of chronic diseases at the same time.

[0062] The multi-source data acquisition module described in claim 2 is activated. Medical staff use this module to input basic information and medication orders for different patients according to their chronic disease type, clarifying the medication requirements for each type of chronic disease patient. Patients input medication feedback data through the patient-side interaction module based on their condition and medication usage. This module uses non-invasive sensing technology to collect real-time medication behavior data for each type of chronic disease patient, collects physiological signs related to the effects of medications for each type of chronic disease at a preset frequency, and simultaneously collects medication environment data for all patients. During the collection process, a dedicated noise reduction formula is automatically run to denoise the raw data, removing interfering data to ensure the accuracy of data for different types of patients. After processing, the data is transmitted to the core computing and analysis module according to patient classification.

[0063] After receiving relevant data from multiple types of chronic disease patients, the core operation and analysis module described in claim 3 automatically activates three sub-units to perform calculations according to patient type. The adherence quantification assessment unit, for different types of chronic disease patients, determines the weight of each indicator using a proprietary weighting formula based on the six core assessment indicators specified in claim 3, and then calculates the comprehensive medication adherence score for each patient using the comprehensive adherence scoring formula, classifying them into corresponding adherence levels. The medication trend prediction unit, based on each patient's historical medication behavior data, uses an improved time-series prediction formula to predict the medication adherence change trend for each patient over a future period, identifying the risk of decreased adherence for various patient types in advance. The reminder strategy optimization unit, combining each patient's current adherence score, trend prediction results, and medication feedback data, generates personalized reminder parameters tailored to each patient using a proprietary reminder strategy optimization formula, and synchronously transmits all calculation results to the intelligent reminder module and the medical staff management module according to patient classification.

[0064] After receiving optimized parameters for various patients transmitted by the core computing and analysis module, the intelligent reminder module described in claim 5 automatically activates three sub-units to perform reminder work according to patient type. The reminder format adaptation unit selects an appropriate reminder format based on each patient's lifestyle and individual circumstances, supporting simultaneous reminders in multiple formats to meet the personalized needs of different patients. The reminder intensity adjustment unit adjusts the reminder intensity and interval for each patient using a reminder intensity grading formula, combined with their compliance level, to ensure targeted reminders. The false reminder avoidance unit uses a false reminder avoidance formula to determine each patient's real-time medication status, effectively avoiding false and repeated reminders for various patients. When each patient's preset medication time is reached, it automatically sends a medication reminder of the corresponding format and intensity until the patient confirms medication.

[0065] Medical staff can log in to the medical staff management module described in claim 6, and classify and manage all patients according to chronic disease type through the patient management unit. They can quickly retrieve and view complete medication records for different types of chronic disease patients, including basic information, medical orders, medication behavior data, physiological sign data, and adherence scores. Through the adherence viewing unit, they can view the real-time adherence score, historical change curve, and scores of various assessment indicators for each patient, clearly understanding the medication adherence of various types of patients. Based on the calculation results of the core calculation and analysis module, combined with the characteristics of each patient's condition and chronic disease type, personalized intervention plans can be formulated through the intervention plan formulation unit and synchronized to the corresponding patient end. Through the batch operation unit, operations such as synchronous reminders and batch distribution of intervention plans to batch patients according to chronic disease type can be performed, greatly improving the management efficiency of multiple types of chronic disease patients in internal medicine. Through the data statistical analysis unit, dedicated statistical formulas are run to calculate the overall adherence of various types of chronic disease patients, as well as the overall mean adherence of all patients, the proportion of each adherence level, etc., generating classified statistical reports, providing comprehensive data support for the mixed management of multiple types of chronic diseases in internal medicine.

[0066] The data storage encryption module described in claim 7 automatically receives all patient data transmitted from each module, classifies and stores patient data according to a hierarchical standard of core data, ordinary data, and temporary data, and classifies and archives patient data according to chronic disease type to ensure standardized data storage and easy retrieval; it performs end-to-end encryption processing on all patients' core data through a dedicated encryption formula, and the encryption key is automatically updated according to a preset period, while performing real-time backup and periodic backup operations, storing backup data on multiple off-site servers to prevent data loss; it cleans up all expired temporary data according to a preset period, compresses redundant data, ensures efficient operation of the storage system, strictly protects the privacy of all patients, and authorizes only medical staff and the corresponding patients to view the relevant data.

[0067] The adherence intervention module described in claim 8 receives the calculation results from the core calculation and analysis module and the intervention plans for various types of patients formulated by the medical staff, and then carries out intervention work according to patient type. The intervention measure matching unit matches and executes corresponding intervention measures for patients with different types of chronic diseases and different adherence levels; the intervention effect tracking unit tracks the implementation effect of the intervention measures in real time for each patient through the intervention effect evaluation formula to determine whether the intervention is effective; based on the intervention effect evaluation results, patient medication feedback, and changes in adherence trends, the intervention plan adjustment unit dynamically adjusts the intervention plan for each patient to ensure that the intervention effect for all types of patients continues to improve and helps all patients improve medication adherence.

[0068] The abnormality warning module described in claim 9 receives in real time medication behavior data and physiological sign correlation data of various patients collected by the multi-source data acquisition module, as well as the calculation results of the core calculation and analysis module, and performs abnormality identification and warning work according to patient type. The abnormality identification unit identifies medication abnormalities, compliance abnormalities, and physiological sign abnormalities of each patient in real time through abnormality identification formula; the warning threshold dynamic adjustment unit dynamically adjusts the warning threshold for each patient based on their disease duration, complication type, and chronic disease type; when any abnormality is identified in any patient, the warning linkage and handling unit immediately sends the strongest warning reminder to the patient and sends detailed warning information to the corresponding medical staff. If no action is taken within a preset time after the warning is issued, the patient's family is automatically contacted, and the entire warning handling process is recorded in detail to form a warning handling file, ensuring that abnormalities of all types of patients are handled in a timely manner.

[0069] The cross-terminal synchronization module described in claim 10 automatically runs a dedicated synchronization formula to synchronize data between all patient terminals, medical staff terminals, and the server in real time, ensuring that the data viewed by medical staff and each patient is up-to-date and that the data across different terminals remains consistent. The system self-test calibration module performs self-tests on the entire platform according to a preset cycle. Through a self-test pass / fail judgment formula, it comprehensively tests the operating status, data transmission status, calculation accuracy, and storage system status of each module. If a self-test failure is found, the system is automatically calibrated immediately through a calibration formula. If calibration fails, the system automatically switches to a backup system to ensure continuous and stable operation of the platform. This allows for the orderly implementation of medication adherence monitoring and reminders for multiple types of chronic internal medicine patients, achieving efficient mixed management of chronic internal medicine patients.

[0070] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A medication adherence monitoring and reminder platform for patients with chronic internal medicine diseases, characterized by: The system comprises a patient-side interaction module, a multi-source data acquisition module, a core computation and analysis module, an intelligent reminder module, a medical staff management module, a data storage encryption module, a compliance intervention module, an anomaly warning module, a cross-terminal synchronization module, and a system self-test calibration module. Each module is deployed in a distributed architecture, achieving bidirectional data interaction through a proprietary encryption protocol. Each module operates independently yet collaboratively, without relying on the hardware interfaces and software algorithms of existing medication monitoring systems. It is specifically adapted to medication scenarios for various chronic diseases in internal medicine (hypertension, diabetes, coronary heart disease, chronic bronchitis, post-stroke sequelae, etc.), enabling full-process, multi-dimensional, high-precision monitoring and personalized reminders of medication compliance. It also constructs a proprietary compliance assessment system and intervention logic. The specific architecture meets the following requirements: data transmission latency between modules ≤50ms, data interaction success rate ≥99.92%, system continuous operation stability ≥99.87%, and all data calculations are based on a proprietary mathematical model, avoiding the single-dimensional monitoring and generic reminder logic of existing technologies.

2. The medication adherence monitoring and reminder platform for patients with chronic internal medicine diseases according to claim 1, characterized in that: The multi-source data acquisition module is used to collect comprehensive medication-related data from patients with chronic internal medicine diseases, avoiding the single data acquisition mode in existing technologies. Specific data collected includes basic patient information, medication orders, real-time medication behavior data, physiological sign correlation data, medication environment data, and medication feedback data. Basic patient information includes age, gender, disease duration, complication type, drug allergy history, liver and kidney function indicators, education level, and medication awareness. Medication orders include drug type, specifications, single dose, frequency, medication time window, treatment course, discontinuation criteria, and dosage adjustment rules. Real-time medication behavior data is collected using non-invasive sensing technology, including actual medication time, actual dosage, number of missed doses, number of incorrect doses, duration of unauthorized discontinuation of medication, and dosage adjustment behavior; physiological sign-related data includes indicators directly related to medication efficacy such as blood pressure, blood glucose, heart rate, blood oxygen saturation, and respiratory rate, collected every 30 minutes, with the frequency increased to once every 5 minutes under abnormal conditions; medication environment data includes ambient temperature and humidity, and light intensity (to avoid misjudgment of medication adherence due to improper drug storage); medication feedback data includes post-medication discomfort symptoms, evaluation of medication convenience, and feedback on factors affecting adherence; A noise reduction algorithm is used during data acquisition. The noise reduction formula is as follows: ,in These are the data values ​​after noise reduction. These are the original collected data values. The sample size for a single data collection group ( ≥30), This is the average of the original data. The noise reduction coefficient (range 0.02-0.08, dynamically adjusted according to data type) ensures that the accuracy of the collected data is ≥99.7%, and the collection method does not affect the patient's normal life, avoiding the problems of existing invasive collection and data omission and miscollection.

3. The medication adherence monitoring and reminder platform for patients with chronic internal medicine diseases according to claim 1, characterized in that: The core computation and analysis module, as the core processing unit of the platform, avoids the generalized computational logic and single evaluation model of existing technologies. It adopts a multi-dimensional fusion computational algorithm, based on the full amount of data collected by the multi-source data acquisition module, to achieve accurate quantitative assessment of patient medication adherence, prediction of medication behavior trends, and dynamic optimization of reminder strategies. Specifically, it includes three sub-units: a adherence quantitative assessment unit, a medication trend prediction unit, and a reminder strategy optimization unit. The adherence quantitative assessment unit constructs a specific adherence assessment index system for chronic internal medicine diseases. The indicators include six core indicators: medication on-time rate, dosage accuracy rate, treatment course completion rate, missed dose frequency, incorrect dose frequency, and duration of unauthorized discontinuation of medication. The weights of each indicator are determined using the analytic hierarchy process (AHP), and the weight calculation formula is as follows: in For the first Weight of each indicator ( =1,2,...,6), For the first A senior internal medicine expert on the first Scoring of each indicator ( ≥15 (score range 1-10), and meets the following conditions The final compliance score calculation formula is as follows: ,in The compliance score is calculated as follows (range 0-100 points). For the first The actual score of each indicator (range 0-100 points); based on the comprehensive score, compliance is divided into 5 levels: high compliance ( ≥90 points), high compliance (80 points ≤ <90 points), moderate compliance (65 points ≤ <80 points), low compliance (50 points ≤ <65 points), low compliance ( <50 points), each level corresponds to a specific intervention threshold and reminder intensity, avoiding the problems of overly coarse compliance grading and inaccurate assessment in existing technologies.

4. The medication adherence monitoring and reminder platform for patients with chronic internal medicine diseases according to claim 3, characterized in that: The medication trend prediction unit employs an improved time-series prediction algorithm to overcome the limitations of existing single-trend prediction models. Based on patients' historical medication behavior data (containing at least 3 months of complete data), it predicts the trend of changes in patients' medication adherence over the next 1-4 weeks. The prediction formula is as follows: ,in For the first Zhou's compliance prediction score For the first Zhou's actual compliance score This is a trend adjustment factor (range 0.05-0.12, dynamically adjusted according to the duration of the disease; the longer the disease duration, the higher the trend adjustment factor). (The smaller the value) To backtrack the number of weeks for historical data ( ≥12), For the first The difference in weekly physiological signs (calculated by comparing current physiological sign data with the average of previous periods). The physiological sign influence coefficient (range 0.10-0.18, dynamically adjusted according to the type of chronic disease, including patients with diabetes and hypertension). (Value ≥ 0.15); prediction error controlled within ±3.5 points, prediction accuracy ≥ 92.3%, capable of identifying the risk of declining adherence in advance, providing data support for the implementation of subsequent intervention measures; the reminder strategy optimization unit dynamically optimizes the reminder parameters based on adherence scores, medication trend prediction results, and patient medication feedback data, with the optimization formula as follows: The optimized reminder lead time (in minutes). The standard reminder time is set in advance (default 30 minutes, which can be adjusted by medical staff). The current overall compliance score, To predict compliance scores, To optimize the coefficient (ranging from 0.03 to 0.07), and ensure that the reminder strategy is highly compatible with patients' medication habits and compliance levels, the problem of fixed reminder strategies and inability to be dynamically adjusted in existing technologies is avoided.

5. The medication adherence monitoring and reminder platform for patients with chronic internal medicine diseases according to claim 1, characterized in that: The intelligent reminder module avoids the problems of single-form reminders, fixed reminder intensity, and high false reminder rate in existing technologies. It adopts a multi-mode integrated reminder method, which combines the patient's compliance level, medication scenario, physiological signs and medication feedback to achieve personalized and hierarchical reminders. Specifically, it includes three sub-units: a reminder form adaptation unit, a reminder intensity adjustment unit, and a false reminder avoidance unit. The reminder formats include voice reminders, text reminders, vibration reminders, light reminders, and reminders linked to the medical staff. Patients or medical staff can choose according to the patient's lifestyle (such as whether they have hearing impairment or are asleep), and multiple forms of reminders can be synchronized simultaneously. The reminder intensity adjustment unit dynamically adjusts the reminder intensity based on the compliance level. The intensity grading formula is as follows: ,in The current alert level (range 1-5). The lowest alert level (Level 1). This is the highest alert level (Level 5). The current overall adherence score is used. The lower the adherence, the higher the reminder intensity, and the reminder interval gradually shortens (30-minute reminder interval for high adherence, 5-minute reminder interval for low adherence, until the patient confirms medication use). The false reminder avoidance unit determines whether the patient has completed medication use by identifying the patient's medication behavior data and physiological sign data, avoiding duplicate and false reminders. The determination formula is: in This indicates that the medication has been taken and no further reminder is needed. This indicates that the medication has not been completed and a reminder is needed. This refers to the actual time of medication administration. For standard medication time, This is the actual dosage. For standard medication dosage, Physiological data at the time of medication administration. , Within the normal range of physiological signs, the false alarm rate should be controlled at ≤0.8%.

6. The medication adherence monitoring and reminder platform for patients with chronic internal medicine diseases according to claim 1, characterized in that: The described medical staff management module avoids the problems of existing medical staff terminals having limited functionality and being unable to achieve precise intervention and batch management. It is specifically adapted to the work scenarios of internal medicine medical staff, enabling full-process management, precise intervention, and batch operation of medication adherence for patients with chronic diseases under their care. Specifically, it includes five sub-units: patient management unit, adherence viewing unit, intervention plan formulation unit, batch operation unit, and data statistical analysis unit. Among them, the patient management unit allows medical staff to classify and manage patients according to dimensions such as chronic disease type, adherence level, disease duration, and ward. Patient information can be added, edited, and deleted, and patients' complete medication records can be viewed (including basic information, medical order information, medication behavior data, physiological sign data, adherence score, intervention records, etc.). It also supports fast retrieval and fuzzy search of patient information, with a retrieval response time of ≤1 second. The adherence monitoring unit allows healthcare professionals to view a single patient's real-time adherence score, historical adherence change curves, specific scores for various assessment indicators, and analysis of the reasons for adherence changes (automatically generated based on medication behavior data and physiological sign data). The intervention plan development unit allows healthcare professionals to develop personalized adherence intervention plans based on patients' adherence levels, medication trends, physiological sign status, and complication status. These plans include medication guidance, psychological counseling suggestions, dosage adjustment reminders, and follow-up plans, and the intervention plans can be automatically synchronized to the patient's end for real-time tracking of intervention effects. The batch operation unit supports healthcare professionals in simultaneously reminding batches of patients, batch distributing intervention plans, and batch developing follow-up plans, improving healthcare management efficiency. The data statistical analysis unit automatically calculates the overall compliance of patients under its jurisdiction. The statistical formulas include: the formula for the overall compliance mean. (in The overall compliance mean. For the first Patient compliance scores, Formula for total number of patients under management and percentage of each compliance level. (in For the first The percentage of each compliance level For the first The number of patients at each compliance level =1,2,...,5), the statistical results are displayed in the form of charts and graphs, and statistical reports can be exported, providing data support for the management of medication for chronic diseases in internal medicine.

7. The medication adherence monitoring and reminder platform for patients with chronic internal medicine diseases according to claim 1, characterized in that: The data storage encryption module avoids the problems of insecure data storage, data redundancy, low query efficiency, and high privacy leakage risks in existing technologies. It adopts a layered encrypted storage architecture combined with a proprietary encryption algorithm to achieve secure storage, efficient querying, and privacy protection for all patient medication-related data. Specifically, it includes four sub-units: a layered data storage unit, a data encryption unit, a data backup unit, and a data cleaning unit. The layered data storage unit divides data into three levels: core data (patient basic information, medical order information, adherence scores), general data (medication behavior data, physiological sign data), and temporary data (medication environment data, temporary medication feedback records). Each level uses a different storage method: core data uses a distributed database, general data uses a relational database, and temporary data uses a cache. The storage capacity is dynamically expandable, supporting a single patient data storage duration of ≥5 years and a data query response time of ≤0.5s. The data encryption unit uses a proprietary symmetric encryption algorithm to encrypt the core data throughout the entire process. The encryption formula is: ,in For encrypted data, For the original core data, Use a dedicated encryption key (automatically generated by the system, updated every 30 days, key length ≥ 256 bits). The encryption offset (values ​​ranging from 100 to 200, randomly generated) is used to decrypt the following formula: To ensure that core data is not leaked or tampered with during storage and transmission; The data backup unit employs a combination of real-time and scheduled backups. Real-time backups are performed every minute, while scheduled backups occur daily at 2:00 AM. Backup data is stored on multiple off-site servers, achieving a backup success rate of ≥99.99%. It supports data recovery from accidental deletion (recovery time ≤10 minutes). The data cleanup unit automatically cleans up expired temporary data (cleanup cycle is 7 days) and compresses redundant data with a compression ratio of ≥80%, ensuring the efficient operation of the storage system. Simultaneously, it strictly adheres to medical data privacy protection regulations, authorizing only medical staff and the patients themselves to view patient data, thus mitigating the risk of privacy leaks.

8. The medication adherence monitoring and reminder platform for patients with chronic internal medicine diseases according to claim 1, characterized in that: The adherence intervention module avoids the problems of single intervention measures, weak targeting, and poor intervention effects in existing technologies. Based on the adherence score, medication trend prediction results, and intervention plans formulated by medical staff obtained from the core computing and analysis module, it realizes personalized, full-process intervention for patients' medication adherence. Specifically, it includes three sub-units: an intervention measure matching unit, an intervention effect tracking unit, and an intervention plan adjustment unit. The intervention measure matching unit matches corresponding intervention measures according to the patient's adherence level, chronic disease type, disease duration, medication knowledge level, and factors affecting adherence. Patients with high adherence only receive medication reminder reinforcement services, while patients with low or no adherence receive comprehensive intervention measures such as medication guidance, psychological counseling, family interaction, and follow-up reminders. The intervention measures can be dynamically adjusted based on patient feedback. The intervention effect tracking unit tracks the implementation effect of the intervention measures in real time, using an intervention effect evaluation formula: ,in The intervention effect rate (range: -100% to +100%). For adherence scores after intervention, For compliance scores before intervention, ≥10% indicates that the intervention is effective, while E<0% indicates that the intervention is ineffective and the intervention plan needs to be adjusted. The intervention program adjustment unit automatically or manually adjusts the intervention program based on intervention effectiveness evaluation results, patient medication feedback, and changes in medication trends. The adjustment principle is: if ≥20%, maintain the current intervention plan; if 10% ≤ If <20%, optimize the frequency of intervention implementation; if If the rate is less than 10%, change the type of intervention to ensure that the intervention effect continues to improve, and ultimately achieve the goal of increasing patient compliance by an average of ≥15%.

9. The medication adherence monitoring and reminder platform for patients with chronic internal medicine diseases according to claim 1, characterized in that: The aforementioned anomaly warning module avoids the problems of untimely warnings, fixed warning thresholds, high false alarm rates, and lack of coordinated response in existing technologies. Based on medication behavior data and physiological sign data collected by the multi-source data acquisition module and the calculation results of the core computation and analysis module, it achieves real-time warnings and coordinated response for medication anomalies, adherence anomalies, and physiological sign anomalies. Specifically, it includes three sub-units: an anomaly identification unit, a dynamic warning threshold adjustment unit, and a warning coordinated response unit. Anomaly types include: missed doses ≥3 consecutive times, dosage deviation ≥10%, unauthorized discontinuation of medication ≥24 hours, a sudden drop in adherence score ≥15 points (within 24 hours), and physiological sign data exceeding the normal range for ≥10 minutes. The anomaly identification unit uses an anomaly identification algorithm with the following formula: in This indicates an anomaly and requires triggering an alert. This indicates no abnormalities and no warning is needed. This represents the number of consecutive missed doses. For the duration of unauthorized discontinuation of medication, For the current compliance score, The adherence score is based on data from 24 hours prior; the warning threshold is dynamically adjusted based on the patient's disease duration, complication type, and medication history data. The adjustment formula is as follows: ,in The optimized warning threshold, The standard warning threshold is... The warning threshold is set at the duration of the illness (in years). The longer the illness duration, the more lenient the warning threshold, to avoid excessive warnings. The warning linkage and response unit immediately sends a warning reminder (at the highest intensity) to the patient's end after the warning is triggered, and at the same time sends warning information (including abnormality type, abnormal data, and patient information) to the corresponding medical staff end. If no action is taken within 30 minutes after the warning is issued, the system will automatically contact the patient's family (with pre-reserved contact information) and generate a warning response record, recording the warning time, abnormality type, response method, and response effect. The false alarm rate is controlled at ≤1.2%, and the warning response time is ≤10s.

10. The medication adherence monitoring and reminder platform for patients with chronic internal medicine diseases according to claim 1, characterized in that: The cross-terminal synchronization module and system self-test calibration module work together to avoid the problems of cross-terminal data asynchrony, system instability, and decreased computational accuracy in existing technologies. The cross-terminal synchronization module supports real-time data synchronization between patient terminals (mobile phones, tablets, smartwatches), medical staff terminals (computers, mobile phones), and the server, using a dedicated synchronization protocol. The synchronization formula is as follows: ,in For the synchronized data, For data from the data source terminal, For the target terminal's data, The synchronization weight is set to 0.8-0.9, with higher priority data source terminals. (The larger the value), the synchronization delay is ≤100ms, and the data synchronization accuracy is ≥99.95%, ensuring that the data viewed by patients and medical staff is consistent; the system self-test calibration module adopts a combination of periodic self-testing and real-time calibration, with a self-testing cycle of once every 2 hours. The self-testing content includes the operating status of each module, data transmission status, calculation accuracy, and storage system status. The formula for judging whether the self-test is qualified is: in Self=0 indicates that the self-test is passed; Self=0 indicates that the self-test is failed. This indicates the module's operating status (1 indicates normal operation, 0 indicates an error). For data transmission delay, This is due to computational precision error. To store remaining capacity; if the self-test fails, an immediate system alarm will be triggered, and automatic calibration will be performed simultaneously. The calibration formula is: Where Cal is the calibrated operational parameter, These are standard operational parameters. To minimize computational accuracy errors, the calibration success rate is ≥99.8%. If calibration fails, the system automatically switches to the backup system to ensure continuous and stable operation of the platform and that the computational accuracy always meets the requirements.

Images