Cardiometabolic disease medication discontinuation risk identification and intervention pathway generation method

Abstract

The application provides a medication discontinuation risk identification and intervention path generation method for cardiovascular and metabolic diseases, relates to the technical field of medical information processing, chronic disease health management and nursing auxiliary decision-making, and comprises the following steps: acquiring and standardizing patient multi-source data; generating behavior, cognition, environment and other characteristics; calculating a medication discontinuation risk score MIR; performing C1 cognitive deficiency type, C2 drug concern type, C3 behavior forgetfulness type, C4 prescription continuation difficulty type, C5 insufficient support type and C6 cost and information interference type six medication discontinuation disorder typing; generating corresponding intervention paths according to the typing; performing minimum intervention priority sorting on the multiple paths; and dynamically updating the risk and the path according to follow-up feedback. The application can accurately locate the medication discontinuation reason, output an individualized intervention scheme, form a dynamic closed-loop management, improve the chronic disease management efficiency and medication continuity of cardiovascular and metabolic diseases, and is suitable for a hospital follow-up system, an internet medical platform and a community chronic disease management platform.

Description

Technical Field

[0001] This invention relates to the fields of medical information processing, chronic disease health management, and nursing auxiliary decision-making technology. In particular, it relates to a method for identifying medication interruption risk, classifying medication interruption barriers, generating intervention pathways, and updating follow-up feedback for patients with cardiovascular metabolic diseases. Specifically, it relates to a method for identifying medication interruption risk and generating intervention pathways for cardiovascular metabolic diseases. Background Technology

[0002] Cardiovascular metabolic diseases refer to a group of interconnected diseases or risk states, including diabetes, dyslipidemia, hypertension, obesity, and atherosclerotic cardiovascular disease. These patients typically require long-term, continuous use of lipid-lowering drugs, hypoglycemic drugs, antihypertensive drugs, or other cardiovascular metabolic medications. The continuity of long-term medication directly impacts disease control and the prevention of cardiovascular events.

[0003] Taking lipid management as an example, both the "Chinese Guidelines for Lipid Management (2023)" and the "Chinese Expert Consensus on Lipid Management in Diabetic Patients (2024 Edition)" emphasize that patients with diabetes and those at high risk of cardiovascular and metabolic disorders should pay attention to long-term management of lipid indicators such as LDL-C. Early-onset type 2 diabetes patients have early disease onset, long disease duration, and a high cumulative risk of cardiovascular events. Their awareness, implementation, and achievement rates in lipid management are insufficient, and the influencing factors not only stem from individual disease awareness and treatment adherence but also involve multiple levels of factors such as family support, communication with healthcare professionals, community resources, and policy guarantees.

[0004] In recent years, digital health, mobile health, and eHealth interventions have been widely used to improve chronic disease management. However, most tools still remain at the level of reminders, check-ins, and general education, lacking the ability to identify the causes of medication interruption and lacking mechanisms for automatically generating intervention pathways based on these classifications.

[0005] Existing chronic disease management systems for cardiovascular and metabolic diseases typically focus on medication reminders, follow-up appointment reminders, indicator recording, and general health education. For example, existing lipid management methods can collect identity information, lipid information, and assessment information, and perform lipid dyslipidemia assessment and ASCVD risk assessment; existing diabetes management systems also include modules for patient information management, doctor-patient interaction, disease assessment, reminders, and wireless transmission. While these technologies can remind patients to "take medication" and "have follow-up appointments," they have the following drawbacks:

[0006] 1. Existing systems typically only identify whether a patient has irregular medication use, delayed follow-up examinations, or abnormal indicators, but cannot further identify the specific reasons for medication discontinuation. Long-term medication discontinuation in patients with cardiovascular and metabolic diseases may be caused by a combination of factors, including insufficient disease awareness, concerns about drug side effects, behavioral amnesia, difficulty in resuming prescriptions, insufficient family support, financial burden, or misleading information online. If the system only outputs "poor adherence" or "high risk," medical staff still need to manually determine the specific source of the obstacle, resulting in low management efficiency.

[0007] 2. Existing systems primarily rely on single reminders or standardized health education as intervention methods, lacking a corresponding mechanism for "disorder classification - intervention path". Different patients experience medication discontinuation for different reasons; using the same education and reminders for all patients can easily lead to inaccurate interventions and fail to effectively address the actual obstacles faced by patients.

[0008] 3. Existing systems primarily focus on disease indicators or risk values ​​themselves, such as dyslipidemia, ASCVD risk, and blood glucose control status, but lack dynamic risk identification methods for long-term medication continuity. Patients with cardiovascular and metabolic diseases often require long-term medication, and a single assessment is insufficient to reflect subsequent changes due to missed doses, discontinuation of medication, failed prescription renewals, or interrupted follow-up examinations.

[0009] 4. The existing system lacks a dynamic update mechanism after follow-up feedback. After receiving reminders, education, prescription renewal assistance, pharmacist consultation, or family collaboration, a patient's risk status may change. If the system cannot recalculate the risk and adjust the path based on the feedback data, it will be difficult to form a closed-loop management system.

[0010] Therefore, current technological focus is shifting from "simple reminders" to "multi-source data for risk identification, obstacle location, generation of personalized intervention paths, and continuous feedback and updates."

[0011] The published patent CN109887575A, "A Method and System for Blood Lipid Management," proposes a structure including a patient module, an acquisition unit, a communication unit, an account generation unit, a blood lipid abnormality assessment unit, and an ASCVD risk assessment unit. The basic steps of this scheme are: collecting patient identity information, blood lipid information, and assessment information; the system judges abnormal blood lipid values; compares current blood lipid values ​​with standard or historical values; performs ASCVD risk assessment based on the assessment information; and displays blood lipid changes and risk warnings to the patient. However, the shortcomings of this type of scheme are: its focus is on judging abnormal blood lipids and assessing ASCVD risk, but it does not address why patients cannot continue taking related medications, why they cannot complete follow-up examinations, or why they stop taking medication or interrupt prescription renewals. Furthermore, it does not establish a correspondence between medication interruption disorder classifications and intervention pathways.

[0012] The published patent CN108511070A, "A Diabetes Patient Assessment and Management System," includes a central processing module, an interface management module, a doctor-patient interaction module, a condition assessment module, a patient information management module, a reminder module, a feedback module, and a wireless transmission module. The basic structure of this solution is as follows: the patient information management module manages basic patient information and disease information; the condition assessment module assesses the patient's physical condition based on indicators; the doctor-patient interaction module is used for consultation and interaction; the reminder module reminds patients of medication, diet, exercise, and follow-up examinations; the feedback module collects patient feedback; and the wireless transmission module transmits information within the system. However, this type of solution has shortcomings: it primarily focuses on comprehensive diabetes assessment and routine reminders. While it can remind patients to take medication on time and in the correct dosage and to undergo relevant examinations, it does not refine the reasons for medication interruption into actionable obstacle classifications, nor does it generate differentiated intervention paths based on different obstacle classifications, and it does not provide a "minimum intervention priority ranking" to address the problem of path selection when multiple obstacles coexist.

[0013] Current chronic disease management often uses questionnaires or scales to assess patient medication adherence, such as asking patients if they have missed doses, discontinued medication on their own, or renewed prescriptions on time. While these methods can determine patient adherence levels, they generally remain at the level of measurement and statistical analysis. Their shortcomings include: only providing adherence results and not automatically generating intervention pathways; often failing to integrate with hospital prescriptions, laboratory tests, follow-up examinations, and other data; not being able to dynamically update risks and pathways based on patient feedback; and being difficult to directly embed as a complete technical solution into chronic disease management systems. Summary of the Invention

[0014] Therefore, the purpose of this invention is to propose a method for identifying medication interruption risk and generating intervention pathways for cardiovascular metabolic diseases. This method can standardize multi-source data from patients with cardiovascular metabolic diseases, calculate medication interruption risk, identify medication interruption disorder types, and generate corresponding intervention pathways based on disorder types. Furthermore, it can dynamically update the risk and pathways based on follow-up feedback. The method further refines "whether there is a risk of medication interruption" into "why there is a risk of medication interruption." It uniformly classifies patients' medication interruption disorders into cognitive deficiency, medication anxiety, behavioral amnesia, difficulty in continuing medication, insufficient support, and cost and information interference types. It automatically generates corresponding intervention pathways based on different disorder types. When multiple disorders coexist, it prioritizes the most needed, feasible, and likely-to-improve medication continuity pathways by sorting them by minimum intervention priority. Finally, it dynamically updates the risk score, disorder type, and intervention pathway based on follow-up feedback.

[0015] This invention provides a method for identifying the risk of medication discontinuation in cardiovascular metabolic diseases and generating intervention pathways, comprising the following steps:

[0016] S1. Obtain multi-source data of patients with cardiovascular and metabolic diseases, including disease information, medication information, medication behavior information, follow-up behavior information, cognitive belief information, support environment information, and external environment information;

[0017] S2. Standardize the multi-source data, including desensitization, encoding, scoring standardization, time window standardization, and anomaly and missing data marking;

[0018] S3. Generate medication behavior characteristics, follow-up behavior characteristics, cognitive belief characteristics, support environment characteristics, and external environment characteristics based on standardized data;

[0019] S4. The medication interruption risk score (MIR) is calculated by combining the generated features. This score is calculated by weighting the risk assessment results of five dimensions: medication behavior, follow-up behavior, cognitive beliefs, supportive environment, and external environment. The individual assessment results of the above five dimensions have been standardized, and the value range is uniformly set to 0 to 1. The larger the value of a single item, the higher the risk of medication interruption for the corresponding dimension. Each dimension is assigned a corresponding weight coefficient, and the sum of all weight coefficients is 1. The final medication interruption risk score (MIR) ranges from 0 to 100. The higher the MIR score, the higher the overall risk of medication interruption for the patient.

[0020] Specifically, the medication interruption risk score MIR is calculated based on the generated features (MIR represents the comprehensive risk of patients missing doses, stopping medication on their own, reducing dosage on their own, changing medication without authorization, failing to renew prescriptions on time, or interrupting follow-up examinations).

[0021]

[0022] Where A represents the risk sub-score for medication behavior; R represents the risk sub-score for follow-up behavior; C represents the risk sub-score for cognitive beliefs; S represents the risk sub-score for the supporting environment; and E represents the risk sub-score for the external environment. , , , , These are the weight coefficients for the sub-ratings of the five dimensions A, R, C, S, and E mentioned above, and they satisfy the following:

[0023] ;

[0024] A, R, C, S, and E are all standardized risk values ​​between 0 and 1, with higher values ​​indicating higher risk in that dimension; MIR values ​​range from 0 to 100, with higher scores indicating a higher risk of medication interruption.

[0025] S5. Classify patients with medication interruption disorder, including C1 cognitive deficiency, C2 medication anxiety, C3 behavioral amnesia, C4 difficulty in continuing medication, C5 lack of support, and C6 cost and information interference.

[0026] S6. Generate corresponding intervention paths based on the obstacle classification results;

[0027] S7. Prioritize multiple intervention pathways by minimum intervention level; dynamically update MIR scores, impairment classifications, and intervention pathways based on follow-up feedback data.

[0028] Furthermore, in step S4, the sub-scores of risk for each dimension are calculated as follows:

[0029] The medication behavior risk sub-score is calculated based on the medication coverage during the patient's observation period. Medication coverage is the proportion of days with medication coverage to the total number of days in the observation period. The medication behavior risk sub-score increases as the medication coverage ratio decreases. When the medication coverage ratio is lower than a preset threshold, it is judged as insufficient medication coverage and is included in the medication interruption risk score.

[0030] The risk sub-score for follow-up behavior is calculated based on the patient's actual follow-up interval and the upper limit of follow-up time recommended by guidelines or consensus. The value of the risk sub-score for follow-up behavior ranges from 0 to 1. When the actual follow-up interval does not exceed the upper limit of the recommended time, the risk sub-score for follow-up behavior is 0. When the actual follow-up interval exceeds the upper limit of the recommended time, the risk sub-score for follow-up behavior increases as the proportion of the time exceeding the upper limit of the recommended time increases, but does not exceed 1.

[0031] Standardized calculations were performed on the corresponding dimension questionnaire scores, which had undergone reliability and validity testing, to obtain cognitive belief risk sub-scores, supporting environment risk sub-scores, and external environment risk sub-scores. For questionnaire items with protective statements, the higher the actual score, the lower the risk, and these scores needed to be converted into risk values ​​positively correlated with the degree of risk. For questionnaire items with risky statements, the higher the actual score, the higher the risk, and these scores were directly converted into risk values ​​positively correlated with the degree of risk. During the conversion process, the theoretical minimum and maximum scores of the corresponding dimensions were used as benchmarks to perform a linear mapping on the actual scores to obtain standardized risk values. The corresponding cognitive belief risk sub-scores, supporting environment risk sub-scores, or external environment risk sub-scores were then calculated based on these risk values.

[0032] Furthermore, the method for classifying patients with medication discontinuation disorders in step S5 includes:

[0033] S51. For each type of obstacle classification, first extract the original fields related to that obstacle, and then uniformly convert the original fields into standardized risk values ​​between 0 and 1. Specifically, for fields describing risk, the higher the actual risk level, the larger the standardized risk value after conversion; for fields describing protection, the higher the actual protection level, the smaller the standardized risk value after conversion. The conversion process uses the theoretical minimum and theoretical maximum scores of the corresponding fields as benchmarks to perform a linear mapping on the actual scores to obtain standardized risk values. If the field is a binary variable, it is directly assigned a value of 0 or 1 based on whether there is risk; if the field is a ternary or multi-category variable, it is converted into a value between 0 and 1 according to the risk level from low to high.

[0034] S52. Calculate the obstacle intensity score for each of the six obstacle types, C1 to C6. The obstacle intensity score is calculated by weighted average of the standardized risk values ​​of each relevant field under the corresponding type. Each field is configured with a corresponding weight coefficient, and the weight coefficient of the field is positively correlated with the degree of influence of the field on the obstacle type. The value range of the six obstacle intensity scores is 0 to 1. The higher the score, the higher the intensity of the medication interruption obstacle of that type. C1 is the cognitive deficiency type, C2 is the medication concern type, C3 is the behavioral amnesia type, C4 is the prescription refill difficulty type, C5 is the lack of support type, and C6 is the cost and information interference type.

[0035] Furthermore, in step S5, after calculating the intensity scores of various obstacles, the identification and judgment of medication interruption obstacles are completed in combination with the preset judgment thresholds; each type of obstacle corresponds to an independent judgment threshold, which can be determined by the Delphi expert inquiry method, the analytic hierarchy process, historical follow-up data, ROC curve analysis, or the chronic disease management parameter table of medical institutions.

[0036] The calculated obstacle intensity score is compared with the corresponding preset threshold. When the obstacle intensity score of a certain type is greater than or equal to the corresponding threshold, the patient is determined to have that type of medication interruption obstacle; when the obstacle intensity score is less than the corresponding threshold, the patient is determined not to have that type of medication interruption obstacle.

[0037] When outputting the classification results, different output formats can be selected according to actual needs: if it is only necessary to determine the patient's main disorder type, the category with the highest disorder intensity score is selected from the six types of disorders and it is determined as the patient's main medication interruption disorder type; if the intensity scores of two or more disorders reach their respective judgment thresholds, the corresponding combinations of disorders are output simultaneously, indicating that the patient has multiple medication interruption disorders at the same time.

[0038] Furthermore, the intervention pathway in step S6 includes: P1 goal cognition pathway, P2 doubt clarification pathway, P3 behavior reminder pathway, P4 prescription renewal guarantee pathway, P5 family collaboration pathway, and P6 accessibility and information correction pathway.

[0039] Methods for generating corresponding intervention paths based on obstacle classification results include:

[0040] S61. Data reception: Obtain the results of the six types of medication interruption obstacles obtained in the aforementioned steps, as well as the obstacle intensity score corresponding to each type of obstacle. The obstacle intensity score ranges from 0 to 1, and the higher the score, the more prominent the corresponding obstacle is.

[0041] S62. Pre-establish the corresponding association rules between the disorder classification and the intervention path, clarify the matching relationship between the six types of medication interruption disorders and the six types of intervention paths, and adopt a one-to-one correspondence basic association form. Determine whether a certain disorder can trigger the corresponding intervention path based on the association rules.

[0042] S63. Combining the obstacle determination results, obstacle intensity scores, and preset association rules, comprehensively calculate the triggering degree of each type of intervention path; if the triggering degree of a certain intervention path is greater than zero, then the generation of that type of intervention path is initiated; if the triggering degree is zero, then the corresponding intervention path is not generated; in the one-to-one correspondence association mode, when it is determined that a patient has a certain type of medication interruption obstacle, the intensity score of that type of obstacle is the triggering basis for the corresponding intervention path, and a matching intervention path is generated accordingly.

[0043] S64. After determining the intervention path to be generated, call the corresponding intervention task package according to the path type.

[0044] Furthermore, step S64, which involves calling the corresponding intervention task package based on the path type, includes:

[0045] P1 Goal Cognition Path: Corresponding to C1 Cognitive Deficiency Type, the task package includes explanation of long-term medication goals, explanation of disease indicator target values, education on the necessity of continuous management, and explanation of the purpose of follow-up examination;

[0046] P2 Concern Clarification Path: Corresponding to C2 Drug Concern Type, the task package includes identification of common adverse drug reactions, consultation reminders when discomfort occurs, and pharmacist or medical staff consultation tasks.

[0047] P3 Behavioral Reminder Path: Corresponds to C3 Behavioral Forgetting Type, the task package includes fixed medication time reminders, medication check-in, and triggering medical staff follow-up for consecutive missed doses;

[0048] P4 Renewal Guarantee Pathway: Corresponding to the C4 Renewal Difficulty Type, the task package includes advance renewal reminders, community renewal prompts, online renewal prompts, and reminders for review and renewal coordination;

[0049] P5 Family Collaboration Path: Corresponding to the C5 insufficient support type, the task package includes family education, joint reminders for family members, reminders for accompanying family members to follow-up visits, and family members' participation in follow-up tasks.

[0050] P6 Accessibility and Information Correction Path: Corresponding to C6 Cost and Information Interference Type, the task package includes cost burden assessment, medical insurance or reimbursement reminders, drug accessibility reminders, correction of network misinformation, and risk warnings of health product substitution.

[0051] Furthermore, the method for prioritizing multiple intervention paths by minimum intervention priority in step S7 includes:

[0052] S71. Screening candidate intervention paths: Based on the obstacle classification results, obstacle intensity scores, and preset obstacle-intervention path association rules, determine whether each intervention path meets the activation conditions; intervention paths that meet the activation conditions are included in the candidate path set, while paths that do not meet the activation conditions are directly excluded.

[0053] S72. Calculate the priority score for each candidate path: For each intervention path in the candidate set, calculate the priority score by considering multiple dimensions. The scoring reference dimensions include the degree of path triggering, the overall risk of medication interruption for patients, the expected improvement effect of the intervention path, and the actual difficulty of path execution. The overall risk of medication interruption is normalized according to the standard range. The lower the difficulty of path execution, the higher the corresponding score ratio.

[0054] S73. Determine the expected utility value of the intervention path: The expected utility value of the intervention path is used to measure the estimated effect of the path in improving medication continuity, reducing the risk of medication interruption, or eliminating barriers to medication use. The utility value ranges from 0 to 1, with higher values ​​indicating more ideal expected improvement. The expected utility value can be calculated based on historical follow-up data, mainly comparing the change in the overall risk of medication interruption before and after the implementation of the same type of intervention path. If the path targets a single medication barrier, the change in the intensity of the corresponding barrier before and after the intervention is compared to determine the actual utility level of the path.

[0055] S74. Determine the implementation burden of the intervention path: The implementation burden of the intervention path is used to characterize the difficulty of implementing the path. The value range is from 0 to 1. The higher the value, the greater the implementation pressure. The implementation burden is comprehensively evaluated by combining four factors: the difficulty of operation for patients, the difficulty of cooperation for family members, the workload of medical staff, and the economic costs incurred. Each factor is assigned a corresponding weight, and the final implementation burden result is obtained by integrating and calculating according to the weight ratio.

[0056] S75. Complete the path sorting and result output: Based on the calculated priority score, the candidate intervention paths are sorted in descending order of score; in actual application, the upper limit of the number of paths that can be executed in a single follow-up can be preset according to the task capacity of the medical institution's follow-up management, and finally, the top-ranked intervention paths are selected as the execution plan for this time and output.

[0057] Furthermore, the method for dynamically updating based on follow-up feedback data in step S7 includes:

[0058] The evaluation cycle is divided into two rounds, with the previous round being designated as the preceding evaluation round and the current round as the current follow-up round. First, various feedback data generated during the current follow-up round are collected and used as the basic input for data updates in this round.

[0059] The collected feedback data undergoes standardized transformation: feedback indicating increased risk is uniformly converted to higher standardized risk values; feedback indicating reduced risk is uniformly converted to lower standardized risk values; if the feedback is continuous data or a scale scoring format, the theoretical highest and lowest scores for the corresponding content are used as reference standards to complete the numerical conversion within the 0 to 1 range; if the feedback is a binary option, it is directly differentiated according to whether a risk has occurred.

[0060] After standardizing the feedback data, the risk sub-scores for the five dimensions of medication behavior, follow-up behavior, cognitive beliefs, supportive environment, and external environment are recalculated. Combined with the preset weights for each dimension, the initial medication interruption risk score for this round is calculated. To avoid significant score deviations caused by fluctuations in single follow-up data, a smooth iteration method is used to determine the final medication interruption risk score for this round. This score is calculated by combining the risk scores from previous rounds with the initial score for this round, according to a preset update coefficient. The update coefficient can be set according to actual needs. When the update coefficient is at its maximum value, the final score is determined entirely using the data from this round. When the update coefficient is at other values ​​within the range, a smooth update is achieved by considering both historical and current data.

[0061] Subsequently, the original obstacle intensity scores for the six types of medication interruption obstacles were recalculated. The score for each obstacle was calculated by weighting the standardized risk values ​​of all fields in the category with the corresponding weights. The final intensity score for each obstacle was updated using a smooth iterative method, combining the obstacle scores from previous rounds, the original score for this round, and the specific update coefficients to complete the calculation. Then, based on the preset judgment thresholds for each type of obstacle and the updated obstacle intensity scores, the medication interruption obstacle type for the patient in this round was re-determined.

[0062] Based on the latest obstacle assessment results and obstacle intensity scores, and referring to the corresponding rules between obstacle classification and intervention paths, the activation conditions of each intervention path are recalculated to select candidate intervention paths that can be executed in this round.

[0063] Finally, the intervention tasks executed in this round are statistically analyzed, the task completion status is calculated, and the number of completed tasks is compared with the total number of tasks allocated by the system to obtain the task completion rate of the corresponding intervention path, thus completing the dynamic update of the entire process.

[0064] This invention does not directly output disease diagnosis conclusions, nor does it directly determine drug types, drug dosages, or treatment plans. Instead, it outputs the risk of medication interruption, disorder classification, and intervention pathways. It can be applied to hospital chronic disease follow-up systems, internet medical platforms, community chronic disease management platforms, and nursing management systems, for medical staff to refer to when conducting chronic disease management and nursing follow-up.

[0065] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the steps of the method for identifying and generating intervention pathways for medication interruption in cardiovascular metabolic diseases as described above.

[0066] The present invention also provides a computer device, the computer device including a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, it implements the steps of the method for identifying and generating intervention pathways for medication interruption in cardiovascular metabolic diseases as described above.

[0067] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0068] The method for identifying medication interruption risks and generating intervention pathways for cardiovascular metabolic diseases provided by this invention, compared to existing systems that only assess dyslipidemia and ASCVD risk, not only focuses on blood lipids or disease indicators but also on the long-term continuity of medication use, and can identify medication interruption risks such as missed doses, discontinuation, failed prescription renewals, and missed follow-up examinations. Compared to existing systems that typically only include modules for patient information management, disease assessment, doctor-patient interaction, and reminders, this invention further refines the reasons for medication interruption into six obstacle classifications, answering "why might a patient interrupt medication," rather than simply reminding "the patient should take medication." Compared to existing medication adherence questionnaires or scales, this invention does not simply output adherence scores but integrates adherence-related data with disease information, follow-up behavior, cognitive beliefs, supportive environment, cost, and informational environment data to form an MIR score, obstacle classification, and intervention pathway. Compared to traditional digital health reminder tools, this invention can automatically generate differentiated paths based on different obstacle classifications. For example, it outputs a target cognitive path for patients with cognitive deficiencies, a path to clarify doubts for patients with medication concerns, a path to ensure medication continuity for patients with difficulty in resuming medication, and a path to accessibility and information correction for patients with cost and information interference. Compared to existing unified chronic disease intervention programs, this invention introduces a minimum intervention priority ranking. When a patient has multiple obstacles, the system can prioritize outputting the most necessary, feasible, and likely path to improve medication continuity, effectively improving the efficiency of medical intervention. Compared to existing one-time assessment systems, this invention dynamically updates risks and paths based on follow-up feedback, forming a continuous closed-loop management system for long-term sustainable chronic disease management. It can be embedded in hospitals, the internet, and community platforms, covering a variety of cardiovascular and metabolic diseases, and has broad application prospects. Attached Figure Description

[0069] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention.

[0070] In the attached diagram:

[0071] Figure 1 This is a diagram of the overall system architecture.

[0072] Figure 2 This is a flowchart of the data processing method according to an embodiment of the present invention;

[0073] Figure 3 This is a diagram showing the mapping relationship between obstacle classification and intervention path in an embodiment of the present invention;

[0074] Figure 4 This is a schematic diagram of the configuration of a computer device according to an embodiment of the present invention. Detailed Implementation

[0075] 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 systems and products consistent with some aspects of this disclosure as detailed in the appended claims.

[0076] The terminology used in this disclosure is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. The singular forms “a,” “the,” and “the” as used in this disclosure and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

[0077] It should be understood that although the terms first, second, third, etc., may be used in this disclosure to describe various information, such information should not be limited to these terms. These terms are used only to distinguish information of the same type from one another. For example, without departing from the scope of this disclosure, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0078] The embodiments of the present invention will be described in further detail below.

[0079] Example

[0080] This invention provides a method for identifying the risk of medication interruption in cardiovascular metabolic diseases and generating an intervention pathway, comprising the following steps;

[0081] S1. Obtain multi-source data of patients with cardiovascular and metabolic diseases, including disease information, medication information, medication behavior information, follow-up behavior information, cognitive belief information, support environment information, and external environment information;

[0082] Specifically, cardiovascular metabolic diseases include diabetes, dyslipidemia, hypertension, obesity, and atherosclerotic cardiovascular diseases.

[0083] S2. Standardize the multi-source data, including desensitization, encoding, scoring standardization, time window standardization, and anomaly and missing data marking;

[0084] S3. Generate medication behavior characteristics, follow-up behavior characteristics, cognitive belief characteristics, support environment characteristics, and external environment characteristics based on standardized data;

[0085] S4. Calculate the Medication Discontinuation Risk Score (MIR) based on the generated features:

[0086]

[0087] Where A is the medication behavior risk sub-score;

[0088] R represents the risk sub-score for the re-examination behavior;

[0089] C represents the cognitive belief risk sub-score;

[0090] S represents the sub-score supporting environmental risk;

[0091] E represents the external environment risk sub-score;

[0092] , , , , These are the weight coefficients for the sub-ratings of the five dimensions A, R, C, S, and E mentioned above, and they satisfy the following:

[0093] ;

[0094] A, R, C, S, and E are all standardized risk values ​​between 0 and 1, with higher values ​​indicating higher risk in that dimension; MIR values ​​range from 0 to 100, with higher scores indicating a higher risk of medication interruption.

[0095] In step S4, the sub-scores of risk for each dimension are calculated as follows:

[0096] The medication behavior risk sub-score is calculated based on the medication coverage during the patient's observation period. Medication coverage is the proportion of days with medication coverage to the total number of days in the observation period. The medication behavior risk sub-score increases as the medication coverage ratio decreases. When the medication coverage ratio is lower than a preset threshold, it is judged as insufficient medication coverage and is included in the medication interruption risk score.

[0097] The risk sub-score for follow-up behavior is calculated based on the patient's actual follow-up interval and the upper limit of follow-up time recommended by guidelines or consensus. The value of the risk sub-score for follow-up behavior ranges from 0 to 1. When the actual follow-up interval does not exceed the upper limit of the recommended time, the risk sub-score for follow-up behavior is 0. When the actual follow-up interval exceeds the upper limit of the recommended time, the risk sub-score for follow-up behavior increases as the proportion of the time exceeding the upper limit of the recommended time increases, but does not exceed 1.

[0098] Standardized calculations were performed on the corresponding dimension questionnaire scores, which had undergone reliability and validity testing, to obtain cognitive belief risk sub-scores, supporting environment risk sub-scores, and external environment risk sub-scores. For questionnaire items with protective statements, the higher the actual score, the lower the risk, and these scores needed to be converted into risk values ​​positively correlated with the degree of risk. For questionnaire items with risky statements, the higher the actual score, the higher the risk, and these scores were directly converted into risk values ​​positively correlated with the degree of risk. During the conversion process, the theoretical minimum and maximum scores of the corresponding dimensions were used as benchmarks to perform a linear mapping on the actual scores to obtain standardized risk values. The corresponding cognitive belief risk sub-scores, supporting environment risk sub-scores, or external environment risk sub-scores were then calculated based on these risk values.

[0099] Specifically, the medication behavior risk sub-score A = 1 - PDC, where PDC is the proportion of days with medication coverage to the total number of days during the observation period, calculated using the following formula:

[0100] PDC = Number of days with drug coverage during the observation period / Total number of days in the observation period;

[0101] When PDC < 80%, it is marked as insufficient medication coverage and included in the medication interruption risk score.

[0102] Specifically, PDC is a commonly used indicator in medication adherence evaluation. PQA-related medication adherence indicators use PDC as the evaluation method and 80% as the commonly used threshold, that is, PDC≥80% indicates that the adherence evaluation standard has been met.

[0103] The risk sub-score R for follow-up behavior is calculated based on the follow-up time window recommended by guidelines or consensus, and the calculation formula is as follows:

[0104] ;

[0105] Where T represents the actual number of days since the most recent review of the relevant indicator; G represents the upper limit of the review time recommended by the guidelines or consensus; and R ranges from 0 to 1.

[0106] Specifically, for lipid-lowering drug management scenarios, the value of G can be determined according to guidelines or consensus: For first-time users of lipid-lowering drugs, blood lipids, liver enzymes and creatine kinase should be rechecked within 4-6 weeks of medication; if blood lipid parameters reach the target value and there are no adverse drug reactions, the frequency of rechecking can be gradually changed to once every 3-6 months; for those who have achieved the target value in the long term, rechecking once a year is sufficient.

[0107] The risk sub-scores C (cognitive belief risk), S (supportive environment risk), and E (external environment risk) were standardized using questionnaire dimension scores that had been tested for reliability and validity. If the corresponding dimension item was a protective statement, such as "understanding long-term medication goals," "understanding the necessity of continued medication," "family support for continued medication," or "adequate medical and nursing follow-up support," it was converted into a risk value using the following formula:

[0108] ;

[0109] If the corresponding dimension item is a risk statement, such as "concern about drug side effects affecting medication adherence", "drug cost burden affecting continued medication use", "inconvenience in renewing prescriptions affecting medication adherence", or "online information affecting confidence in medication use", then it is converted into a risk value using the following formula:

[0110] ;

[0111] Where X represents the actual score of the dimension, Xmin represents the theoretical minimum score of the dimension, and Xmax represents the theoretical maximum score of the dimension;

[0112] Risk value Perform standardized calculations to obtain C, S, or E.

[0113] If a dimension contains multiple items, the average of the standardized risk values ​​of each item within that dimension, the weighted average, or the dimension score determined by expert consensus can be used as the sub-score for that dimension.

[0114] Specifically, the weight coefficients αA, αR, αC, αS, and αE are determined by at least one of the following methods: Delphi expert consultation method, analytic hierarchy process, training with historical follow-up data, logistic regression, Cox regression, random forest, gradient boosting model, or other verifiable data modeling methods.

[0115] The risk level classification of MIR is determined by at least one of the following methods: determining an initial threshold based on expert consensus; determining the optimal cutoff value based on historical follow-up data, with the actual occurrence of medication interruption events as the outcome variable, and using ROC curves; or classifying patients into four management levels—low risk, medium risk, high risk, and very high risk—according to the resource allocation needs of medical institutions for chronic disease management.

[0116] Through the above-mentioned medication interruption risk scoring method, this invention can integrate multiple types of data such as prescription coverage, follow-up behavior, cognitive beliefs, supportive environment and external environment, and can adjust the weight coefficients and risk level thresholds according to different disease types, drug types, medical institution follow-up standards and historical data.

[0117] S5. Classify patients with medication interruption disorder, including C1 cognitive deficiency, C2 medication anxiety, C3 behavioral amnesia, C4 difficulty in continuing medication, C5 lack of support, and C6 cost and information interference.

[0118] Methods for classifying patients with medication discontinuation disorders include:

[0119] S51. Standardization of Original Fields: For each obstacle classification, firstly, extract the original fields related to that obstacle and uniformly convert the original fields into standardized risk values ​​between 0 and 1. Specifically, for fields describing risk, the higher the actual risk level, the larger the standardized risk value after conversion; for fields describing protection, the higher the actual protection level, the smaller the standardized risk value after conversion. The conversion process uses the theoretical minimum and theoretical maximum scores of the corresponding field as benchmarks to perform a linear mapping on the actual scores to obtain standardized risk values. If the field is a binary variable, it is directly assigned a value of 0 or 1 based on whether there is risk; if the field is a ternary or multi-category variable, it is converted into a value between 0 and 1 according to the risk level from low to high.

[0120] Specifically, for each type of obstacle classification, the original fields related to the obstacle are first extracted, and the original fields are uniformly converted into standardized risk values ​​between 0 and 1;

[0121] If the field expresses a risk, such as "concern about drug side effects affecting medication adherence", "frequently forgetting to take medication", "inconvenience in renewing prescriptions affecting medication adherence", or "medication costs affecting medication use", then convert it using the following formula:

[0122] ;

[0123] If the field is a protective statement, such as "Knowing the long-term medication goal", "Family support for continued medication", "Sufficient explanation from medical staff", or "Convenient prescription renewal channels", then it should be converted using the following formula:

[0124] ;

[0125] in, The standardized risk value of the k-th field in the j-th obstacle classification;

[0126] This represents the actual score of the k-th field in the j-th obstacle classification.

[0127] This is the theoretical minimum score for this field;

[0128] This represents the theoretically highest score for this field.

[0129] j represents the obstacle classification category, j=1, 2, 3, 4, 5, 6, corresponding to C1 to C6 respectively;

[0130] k represents the specific field number under this obstacle classification.

[0131] If the field is a binary variable, it can be directly assigned a value: no risk = 0, risky = 1; if the field is a ternary or multi-category variable, it is converted to a value between 0 and 1 according to the level of risk from low to high, for example, no impact = 0, slight impact = 0.5, significant impact = 1.

[0132] S52. Calculate the intensity scores for the six types of barriers: Calculate the intensity scores for the six barrier types C1 to C6 respectively. The intensity scores are calculated by weighted average of the standardized risk values ​​of each relevant field under the corresponding type. Each field is configured with a corresponding weight coefficient, and the weight coefficient of the field is positively correlated with the degree of influence of the field on the barrier type. The intensity scores for the six types of barriers range from 0 to 1. The higher the score, the higher the intensity of the type of medication interruption barrier. C1 is the cognitive deficiency type, C2 is the medication concern type, C3 is the behavioral amnesia type, C4 is the prescription refill difficulty type, C5 is the lack of support type, and C6 is the cost and information interference type.

[0133] Specifically, obstacle intensity scores for the six obstacle types, from C1 to C6, are calculated and denoted as B1B_1B1 to B6B_6B6. The general calculation formula is as follows:

[0134] ;

[0135] in, The severity score for the type j medication interruption disorder is given, with a value ranging from 0 to 1.

[0136] The standardized risk value of the k-th field in the j-th obstacle classification;

[0137] The weight coefficient of the k-th field in the j-th obstacle classification;

[0138] The number of fields included in the j-th type of obstacle classification;

[0139] When j=1, it corresponds to the C1 cognitive deficiency type;

[0140] When j=2, it corresponds to the C2 drug-concerned type;

[0141] When j=3, the corresponding behavior is C3, which is the forgetting type.

[0142] When j=4, it corresponds to the C4 continuation of the difficult type;

[0143] When j=5, it corresponds to the C5 under-support type;

[0144] When j=6, it corresponds to the C6 cost and information interference type.

[0145] Specifically, field weights It can be determined through the Delphi expert inquiry method, the analytic hierarchy process, the results of questionnaire reliability and validity testing, the training results of historical follow-up data, or the chronic disease management parameter table validated by medical institutions.

[0146] C1 cognitive impairment is used to identify the risk of medication discontinuation due to a lack of understanding of disease management goals, the significance of long-term medication, or the necessity of follow-up examinations. Its impairment intensity score can be expressed as:

[0147] ;

[0148] in, Risk value for insufficient awareness of long-term medication goals;

[0149] Risk value for insufficient understanding of the necessity of continued medication;

[0150] This represents the risk value for discontinuing medication simply because there are no symptoms.

[0151] The risk value is the value at which medication can be discontinued if the indicators are considered normal.

[0152] The weights are assigned to the fields mentioned above.

[0153] The C2 drug concern type is used to identify the risk of patients discontinuing medication due to concerns about side effects, long-term medication safety, or drug dependence. Its obstacle strength score can be expressed as:

[0154] ;

[0155] in, To mitigate the risk of drug side effects;

[0156] To mitigate the risk of drug-induced liver and kidney damage;

[0157] Concerned about the risk of long-term medication dependence or burden;

[0158] Risk values ​​for patients who tend to stop or reduce medication on their own after experiencing discomfort;

[0159] Weights corresponding to the above fields

[0160] C3 behavioral amnesia is used to identify the risk of missed doses or interrupted follow-up appointments due to busy work schedules, irregular lifestyles, lack of reminder tools, or failure to establish a consistent medication routine. Its impairment severity score can be expressed as:

[0161] ;

[0162] in, This represents the risk value for missing a dose due to forgetting.

[0163] The risk value for missing a dose due to busy work or irregular schedule;

[0164] for Risk value of lacking medication reminder tools;

[0165] Risk values ​​for those who have not established a fixed medication schedule or medication habit;

[0166] The weights are assigned to the fields mentioned above.

[0167] The C4 prescription refill difficulty level is used to identify the risk of medication interruption due to inconvenience in follow-up visits, complex refill procedures, medication running out before timely refill, or insufficient community or online refill channels. Its obstacle intensity score can be expressed as:

[0168] ;

[0169] in, This represents the risk value of insufficient remaining medication and failure to refill the prescription.

[0170] The complexity of the follow-up process poses a risk.

[0171] Risk value for inconvenience in follow-up visits;

[0172] Risk of insufficient community-based renewal channels;

[0173] is the risk value of unfamiliar or inaccessible Internet prescription renewal channels;

[0174] are the corresponding weights of the above fields.

[0175] Wherein, can be calculated according to the coverage days of remaining medication, and is calculated by the following formula:

[0176] ;

[0177] Wherein, D represents the number of days that the current remaining medication can cover;

[0178] G represents the preset advance days for safe prescription renewal;

[0179] When D≥G: , indicating that there is currently no risk of prescription renewal interruption;

[0180] When D<G: it increases as D decreases, indicating that the risk of prescription renewal interruption increases.

[0181] The specific value of G can be determined according to the prescription management rules of medical institutions, types of drugs, follow-up cycles or expert consensus.

[0182] C5 insufficient support type is used to identify the patient's risk of medication interruption due to insufficient family reminder, insufficient family support for follow-up visits or medication, insufficient medical staff follow-up, and insufficient doctor-patient communication. Its obstacle intensity score can be expressed as:

[0183] ;

[0184] Wherein, represents the risk value of family reminder;

[0185] is the risk value of family not supporting continuous medication or follow-up visits;

[0186] is the risk value of insufficient medical staff follow-up;

[0187] is the risk value of insufficient doctor-patient communication;

[0188] are the corresponding weights of the above fields.

[0189] C6 cost and information interference type is used to identify the patient's risk of medication interruption due to drug cost burden, inconvenient medical insurance or reimbursement, insufficient drug accessibility, the influence of wrong online information, propaganda for substituting health products and other factors. Its obstacle intensity score can be expressed as:

[0190] ;

[0191] in, Risk value for drug cost burden;

[0192] Risk value related to inconvenience in medical insurance or reimbursement;

[0193] Risk value for drug shortages, medication changes, or insufficient drug accessibility;

[0194] Risk value for the impact of online information on medication information;

[0195] Promoting health supplements as alternatives to drugs impacts risk assessment.

[0196] The weights are assigned to the fields mentioned above.

[0197] In step S5, after calculating the intensity scores of various obstacles, the identification and judgment of medication interruption obstacles are completed in combination with the preset judgment threshold. Each type of obstacle corresponds to an independent judgment threshold, which can be determined by Delphi expert inquiry method, analytic hierarchy process, historical follow-up data, ROC curve analysis or chronic disease management parameter table of medical institutions.

[0198] The calculated obstacle intensity score is compared with the corresponding preset threshold. When the obstacle intensity score of a certain type is greater than or equal to the corresponding threshold, the patient is determined to have that type of medication interruption obstacle; when the obstacle intensity score is less than the corresponding threshold, the patient is determined not to have that type of medication interruption obstacle.

[0199] When outputting the classification results, different output formats can be selected according to actual needs: if it is only necessary to determine the patient's main disorder type, the category with the highest disorder intensity score is selected from the six types of disorders and it is determined as the patient's main medication interruption disorder type; if the intensity scores of two or more disorders reach their respective judgment thresholds, the corresponding combinations of disorders are output simultaneously, indicating that the patient has multiple medication interruption disorders at the same time.

[0200] Specifically, in step S5, obstacle classification is determined by calculating the obstacle intensity score Bj for each type and comparing it with a preset threshold θj. If Bj ≥ θj, then the existence of that type of obstacle is determined. This includes:

[0201] After calculating Bj ( After that, the patient's disorder type is determined according to the preset classification threshold. The determination formula is as follows:

[0202] ;

[0203] in, This represents the classification result for the j-th type of obstacle. This indicates that the patient has a discontinuation disorder for this type of medication. This indicates that the patient has not met the criteria for determining this type of medication discontinuation disorder;

[0204] The obstacle intensity score for the j-th obstacle type;

[0205] is the threshold for classifying the j-th type of obstacle.

[0206] It can be determined through the Delphi expert inquiry method, the analytic hierarchy process, historical follow-up data, ROC curve analysis, or the chronic disease management parameter table of medical institutions.

[0207] If only the main obstacle fractals need to be output, then select... The largest obstacle type is used as the primary obstacle type:

[0208] ;

[0209] in, Classify the patient's primary medication discontinuation disorder;

[0210] This means selecting the option that makes j = 1 to 6. The obstacle classification category with the maximum value.

[0211] If multiple Simultaneously reaching their respective thresholds Then the output will be a combination of multiple obstacle types. For example, if The output will be a combination of C2 drug-related anxiety type, C3 behavioral amnesia type, and C4 prescription continuation difficulty type.

[0212] In this embodiment of the invention, the classification of medication interruption disorder is not based on a single problem. Instead, relevant fields obtained from the patient, medical staff, and system data sources are first converted into standardized risk values. Then, the disorder intensity scores of the six disorder classifications are calculated respectively. Finally, the primary disorder classification or multiple disorder classification of the patient is determined based on the disorder intensity scores.

[0213] Preferably, to avoid outputting too many intervention pathways in low-risk patients, the system can link the barrier intensity score with the medication discontinuation risk score (MIR). For each barrier type, a barrier priority identification value can be further calculated:

[0214] ;

[0215] in, This is the priority identification value for the j-th type of obstacle classification;

[0216] The obstacle intensity score for the j-th obstacle type;

[0217] Score the risk of medication interruption.

[0218] When MIR is high and a certain type When the threshold is high, these types of obstacles are given higher priority in the subsequent intervention path generation step. This linkage mechanism can be used for subsequent minimum intervention priority ranking.

[0219] Through the above steps, six obstacle intensity scores B1 can be obtained respectively. , , , , The system outputs the medication interruption disorder classification results for the target subjects based on threshold determination, maximum value determination, or multi-label determination. The classification results may include a single primary disorder classification or a combination of multiple disorder classifications, and may be linked with the MIR score as input for subsequent intervention path generation and minimum intervention priority ranking.

[0220] S6. Generate corresponding intervention paths based on the obstacle classification results;

[0221] The intervention pathway in step S6 includes: P1 goal cognition pathway, P2 doubt clarification pathway, P3 behavior reminder pathway, P4 prescription renewal guarantee pathway, P5 family collaboration pathway, and P6 accessibility and information correction pathway.

[0222] Methods for generating corresponding intervention paths based on obstacle classification results include:

[0223] S61. Data reception: Obtain the results of the six types of medication interruption obstacles obtained in the aforementioned steps, as well as the obstacle intensity score corresponding to each type of obstacle. The obstacle intensity score ranges from 0 to 1, and the higher the score, the more prominent the corresponding obstacle is.

[0224] Specifically, it receives the six obstacle classification results and obstacle strength scores output from the previous step:

[0225] ;

[0226] in, This represents the classification result of the j-th type of obstacle. This indicates that the patient has this type of disorder. This indicates that the patient does not meet the criteria for this type of disorder.

[0227] Simultaneously receives six types of obstacle intensity scores:

[0228] ;

[0229] in, The value represents the obstacle intensity score for the j-th type of obstacle, ranging from 0 to 1. A higher value indicates a more pronounced obstacle of that type.

[0230] S62. Pre-establish the corresponding association rules between the disorder classification and the intervention path, clarify the matching relationship between the six types of medication interruption disorders and the six types of intervention paths, and adopt a one-to-one correspondence basic association form. Determine whether a certain disorder can trigger the corresponding intervention path based on the association rules.

[0231] Specifically, a mapping matrix between obstacle classification and intervention path is established. :

[0232] ;

[0233] in, This represents the mapping relationship between the j-th type of obstacle classification and the i-th type of intervention path; This indicates that the j-th type of obstacle classification can trigger the i-th type of intervention path; This indicates that the j-th type of obstacle classification does not trigger the i-th type of intervention path; These correspond to six intervention pathways, from P1 to P6. These correspond to six types of obstacles, from C1 to C6.

[0234] The obstacle classification and intervention path are based on a one-to-one mapping relationship; an example of the mapping matrix is ​​shown below:

[0235] ;

[0236] Specifically, the mapping relationship between the disability type and the intervention path is as follows: C1 Cognitive Deficiency Type corresponds to P1 Goal Cognition Path; C2 Medication Concern Type corresponds to P2 Doubt Clarification Path; C3 Behavioral Forgetting Type corresponds to P3 Behavioral Reminder Path; C4 Prescription Renewal Difficulty Type corresponds to P4 Prescription Renewal Guarantee Path; C5 Insufficient Support Type corresponds to P5 Family Collaboration Path; C6 Cost and Information Interference Type corresponds to P6 Accessibility and Information Correction Path.

[0237] The above basic mapping relationship is derived from the logical correspondence between disorder types and intervention goals, and does not involve drug prescriptions, drug dosage adjustments, or disease diagnosis conclusions.

[0238] S63. Combining the obstacle determination results, obstacle intensity scores, and preset association rules, the triggering degree of each type of intervention path is comprehensively calculated. If the triggering degree of a certain intervention path is greater than zero, the generation of that type of intervention path is initiated; if the triggering degree is zero, the corresponding intervention path is not generated. In the one-to-one correspondence association mode, when a patient is determined to have a certain type of medication interruption obstacle, the intensity score of that type of obstacle is the triggering basis for the corresponding intervention path, and a matching intervention path is generated accordingly.

[0239] Specifically, the intervention path trigger value is calculated based on the mapping matrix M and the obstacle classification results. Obstacle Intensity Score Calculate the trigger value for the i-th type of intervention path. The calculation formula is:

[0240] ;

[0241] in, : The trigger value for the i-th type of intervention path;

[0242] This represents the mapping relationship between the j-th type of obstacle classification and the i-th type of intervention path;

[0243] This represents the classification result for the j-th type of obstacle.

[0244] The obstacle intensity score for the j-th obstacle type;

[0245] i = 1, 2, 3, 4, 5, 6, corresponding to P1 to P6 respectively;

[0246] j = 1, 2, 3, 4, 5, 6, corresponding to C1 to C6 respectively;

[0247] when When, generate the i-th type of intervention path; when When this occurs, the i-th type of intervention path is not generated.

[0248] Under a one-to-one basic mapping relationship, if a patient exhibits C2 drug-related concerns, that is... And its obstacle strength score is ,but The trigger value for the clarification path is:

[0249] ;

[0250] At this point, the P2 clarification path is generated;

[0251] S64. After determining the intervention path to be generated, call the corresponding intervention task package according to the path type.

[0252] Specifically, when a patient has multiple disorder types simultaneously, multiple intervention pathways are triggered concurrently; for example, if a patient simultaneously has C2 medication-concern type, C3 behavioral amnesia type, and C4 prescription refill difficulty type, then: It can generate P2 clarification path, P3 behavioral reminder path, and P4 continuation guarantee path respectively. If it is necessary to further determine the priority of execution, the trigger values ​​of each path can be set. The minimum intervention priority ranking steps are input, and the subsequent priority ranking module ranks them by combining MIR scores, expected path utility, and path implementation burden.

[0253] An extended mapping matrix can be used to connect obstacle classifications with intervention pathways across different medical institutions, disease types, and medication management scenarios. For example, the C6 cost and information interference type, in addition to triggering the P6 accessibility and information correction pathway, can also trigger the P4 prescription renewal pathway after expert consensus confirmation; the C5 insufficient support type, in addition to triggering the P5 family collaboration pathway, can also trigger the P3 behavioral reminder pathway. In this case, the corresponding position in the mapping matrix M... It can be adjusted from 0 to 1, or set to a correlation strength value between 0 and 1. This adjustment can be determined through the Delphi expert consultation method, the analytic hierarchy process, historical follow-up data analysis, or the chronic disease management parameter table of medical institutions.

[0254] like Using a correlation strength value between 0 and 1, then The calculation is still performed using the following formula:

[0255] ;

[0256] at this time The larger the value, the stronger the contribution of the j-th type of obstacle classification to the triggering of the i-th type of intervention path.

[0257] Using the mapping method described above, trigger values ​​for each intervention path can be calculated based on the medication interruption disorder classification results, disorder intensity scores, and a preset mapping matrix, and corresponding intervention paths can be automatically generated. For patients with a single disorder classification, a basic intervention path corresponding to that classification can be generated; for patients with multiple disorder classifications, multiple intervention paths can be generated simultaneously and input into the subsequent minimum intervention priority ranking step to determine the priority execution order. The mapping matrix can be adjusted based on expert consensus, historical follow-up data, or different disease management scenarios, thereby improving the adaptability of the generated intervention paths.

[0258] In this invention, the mapping between obstacle classification and intervention path is not completed through manual temporary judgment, but is automatically generated by the system based on the obstacle classification judgment result, obstacle intensity score and preset mapping matrix.

[0259] Step S64, which involves calling the corresponding intervention task package based on the path type, includes:

[0260] P1 Goal Cognition Path: Corresponding to C1 Cognitive Deficiency Type, the task package includes explanation of long-term medication goals, explanation of disease indicator target values, education on the necessity of continuous management, and explanation of the purpose of follow-up examinations.

[0261] P2 Concern Clarification Path: Corresponding to C2 Drug Concern Type, the task package includes identification of common adverse drug reactions, consultation reminders when discomfort occurs, and pharmacist or medical staff consultation tasks.

[0262] P3 Behavioral Reminder Path: Corresponds to C3 Behavioral Forgetting Type. The task package includes fixed medication time reminders, medication check-in, and triggering medical staff follow-up for consecutive missed doses.

[0263] P4 Renewal Guarantee Pathway: Corresponding to the C4 Renewal Difficulty type, the task package includes advance renewal reminders, community renewal prompts, online renewal prompts, and reminders for review and renewal coordination.

[0264] P5 Family Collaboration Path: Corresponding to the C5 insufficient support type, the task package includes family education, joint reminders for family members, reminders for accompanying family members to follow-up visits, and family participation in follow-up tasks.

[0265] P6 Accessibility and Information Correction Path: Corresponding to C6 Cost and Information Interference Type, the task package includes cost burden assessment, medical insurance or reimbursement reminders, drug accessibility reminders, correction of network misinformation, and risk warnings of health product substitution.

[0266] The generated intervention pathways are used for chronic disease follow-up, nursing management and health education reference, and do not directly constitute disease diagnosis conclusions, drug prescriptions, drug dosage adjustments or physician treatment decisions.

[0267] S7. Prioritize multiple intervention pathways by minimum intervention level; dynamically update MIR scores, impairment classifications, and intervention pathways based on follow-up feedback data.

[0268] Methods for prioritizing the least-intervention pathways among multiple intervention pathways include:

[0269] S71. Screening Candidate Intervention Paths: Combining obstacle classification results, obstacle intensity scores, and pre-defined obstacle-intervention path association rules, determine whether each intervention path meets the activation conditions. Intervention paths that meet the activation conditions are included in the candidate path set, while paths that do not meet the activation conditions are directly excluded.

[0270] Specifically, based on the mapping results between obstacle classification and intervention pathways, the trigger value for each type of intervention pathway is calculated:

[0271] ;

[0272] in, This is the trigger value for the i-th type of intervention path;

[0273] This represents the mapping relationship between the j-th type of obstacle classification and the i-th type of intervention path;

[0274] This represents the classification result for the j-th type of obstacle. =1 indicates the existence of this type of obstacle. =0 indicates that this type of obstacle does not exist;

[0275] The obstacle intensity score for the j-th obstacle type;

[0276] i = 1, 2, 3, 4, 5, 6, corresponding to the six intervention pathways from P1 to P6 respectively;

[0277] j=1, 2, 3, 4, 5, 6, corresponding to the six obstacle classifications from C1 to C6 respectively.

[0278] when When >0, the i-th type of intervention path enters the candidate path set; when When =0, the i-th type of intervention path does not enter the candidate path set;

[0279] The candidate path set is represented as:

[0280] ;

[0281] in, This represents the set of candidate intervention paths that need to be included in the subsequent ranking.

[0282] S72. Calculate the priority score for each candidate path: For each intervention path in the candidate set, calculate the priority score by considering multiple dimensions. The scoring reference dimensions include the degree of path triggering, the overall risk of patient medication interruption, the expected improvement effect of the intervention path, and the actual difficulty of path implementation. The overall risk of medication interruption is normalized according to a standard range. The lower the difficulty of path implementation, the higher the corresponding score weight.

[0283] Specifically, for the candidate path set For each intervention path in the data, calculate its path priority score:

[0284]

[0285] in, The priority score for the i-th type of intervention path;

[0286] This is the trigger value for the i-th type of intervention path, reflecting the degree of existence and intensity of the obstacle corresponding to this path;

[0287] The risk score for medication interruption is calculated, with a range of 0-100 points.

[0288] To mitigate the overall risk of medication interruption after standardization;

[0289] is the expected utility value of the i-th type of intervention path, with a value range of 0-1;

[0290] The implementation burden value for the i-th type of intervention path is 0-1;

[0291] This indicates the feasibility modification item; a higher value indicates a lower implementation burden and better feasibility.

[0292] The priority scoring formula for this pathway indicates that the more obvious the obstacles to the pathway, the higher the overall risk of medication interruption for patients, the higher the expected utility of the pathway, and the lower the implementation burden, the higher the priority score of the pathway.

[0293] S73. Determine the expected utility value of the path. The expected utility value of an intervention pathway measures its anticipated effectiveness in improving medication continuity, reducing the risk of medication interruption, or eliminating barriers to medication use. The utility value ranges from 0 to 1, with higher values ​​indicating more ideal expected improvements. The expected utility value can be calculated based on historical follow-up data, primarily comparing the change in the overall risk of medication interruption before and after the implementation of the same type of intervention pathway. If the pathway targets a single medication barrier, the change in the intensity of the corresponding barrier before and after the intervention is compared to determine the actual utility level of the pathway.

[0294] Specifically, the expected utility value of the path is used. This represents the expected effect of the i-th type of intervention pathway on improving medication continuity, reducing the risk of medication interruption, or improving the corresponding disorder classification. Its value ranges from 0 to 1, and the higher the value, the better the expected effect of the pathway.

[0295] In one implementation, Calculated from historical follow-up data:

[0296]

[0297] in, To assess the risk of medication interruption before implementing the i-th type of intervention pathway;

[0298] The risk score for medication interruption after implementing the i-th type of intervention pathway;

[0299] This represents the relative utility of this pathway in reducing the risk of medication interruption.

[0300] If an intervention pathway is primarily used to improve a specific type of obstacle, the calculation is also based on changes in obstacle intensity scores:

[0301] ;

[0302] in, : Obstacle strength score of the j-th type of obstacle before the implementation of the intervention path;

[0303] The obstacle intensity score for the j-th obstacle type after the intervention path is implemented;

[0304] Optionally, the expected utility value of the path The determination can be made using either an expert consensus method or a data-driven approach. Expert consensus methods include the Delphi expert inquiry method, the analytic hierarchy process (AHP), or expert consensus on chronic disease management in medical institutions. Data-driven methods include calculating changes in MIR scores, impairment severity scores, or medication interruption rates before and after intervention based on historical follow-up data, and then using this information to determine the appropriate interventions. Update.

[0305] S74. Determine the path implementation burden value The implementation burden of an intervention pathway is used to characterize the ease or difficulty of implementing the pathway, with a value ranging from 0 to 1. A higher value indicates greater implementation pressure. The implementation burden is comprehensively assessed by combining four factors: the difficulty of patient operation, the difficulty of family cooperation, the workload of medical staff, and the economic costs incurred. Each factor is assigned a corresponding weight, and the final implementation burden result is obtained by integrating and calculating according to the weight ratio.

[0306] Specifically, the burden value is implemented using a path. This indicates the burden of the i-th type of intervention path in actual implementation, with a value ranging from 0 to 1. The higher the value, the greater the implementation burden.

[0307] The implementation burden is calculated by comprehensively considering factors such as the patient's operational burden, the family's involvement burden, the medical and nursing workload burden, and the economic cost burden.

[0308] ;

[0309] in, Let be the implementation burden value of the i-th type of intervention path;

[0310] Reduce the burden on patients, such as whether they need to frequently clock in, fill out records, or make appointments proactively;

[0311] Consider the burden on family members, such as whether family members need to help with reminders, accompany patients to medical appointments, or participate in education.

[0312] To reduce the workload of healthcare professionals, such as whether nurses need to make telephone follow-ups, pharmacists need to consult, or doctors need to confirm;

[0313] Economic cost burden, such as whether it increases the patient's transportation, follow-up visits, consultations, or medication-related costs;

[0314] These are the weighting coefficients for the aforementioned burden factors.

[0315] Preferably, to The data can be converted into a standardized value between 0 and 1 through questionnaires, medical records, platform task completion data, or institutional management experience. to It can be determined through expert consensus, the analytic hierarchy process (AHP), or historical implementation data.

[0316] S75. Complete Path Ranking and Result Output: Based on the calculated priority scores, rank the candidate intervention paths in descending order of score. In practical applications, the maximum number of paths that can be executed in a single follow-up visit can be pre-set according to the task capacity of the medical institution's follow-up management. Finally, select the top-ranked intervention paths as the execution plan for this visit and output them.

[0317] Specifically, the candidate path set is calculated. Priority score for each path Then, according to Sort from highest to lowest:

[0318] ;

[0319] in, This is the sorted priority sequence of intervention paths;

[0320] Indicates according to Sort by high to low;

[0321] This indicates that only paths in the candidate path set are sorted.

[0322] The intervention paths are output first based on their ranking. If the medical institution wants to control the number of tasks in a single follow-up visit, it can set a maximum number of output paths K, outputting the top K highest priority paths.

[0323] ;

[0324] in, This is the final set of intervention paths output.

[0325] This indicates selecting the top K intervention paths from the sorted results;

[0326] This represents the maximum number of intervention paths that can be output within a single follow-up visit or management cycle. The appropriate level can be determined based on the medical institution's follow-up capabilities, the patient's risk level, the allocation of medical and nursing resources, or expert consensus.

[0327] Preferably, when a patient has multiple impairment subtypes, the minimum intervention combination is determined based on the degree of impairment coverage. Let S be the set of intervention paths ultimately selected, then the coverage of the patient's impairment by this set can be expressed as:

[0328] ;

[0329] in, Let S be the coverage of the path set S on patient medication interruption barriers;

[0330] The final set of intervention paths to be selected for output;

[0331] This represents the classification result for the j-th type of obstacle.

[0332] The obstacle intensity score for the j-th obstacle type;

[0333] This is an indicator variable for whether the path set S covers the j-th type of obstacle classification;

[0334] When the path set When there exists at least one path that can cover the j-th type of obstacle, =1;

[0335] When the path set When there is no path covering the j-th type of obstacle, =0;

[0336] Preferably, intervention pathways are added sequentially from high to low according to the pathway priority score, until the coverage reaches the preset barrier coverage threshold η, which can be determined by expert consensus, medical institution management requirements, or historical data.

[0337] This method allows for prioritizing coverage of key obstacle classifications without requiring excessive intervention tasks to be output at once, thus achieving the goal of "minimal intervention".

[0338] For example, after the aforementioned obstacle classification steps, a target individual might be simultaneously identified as C2 medication-related, C3 behavioral amnesia, and C4 prescription refill difficulties, i.e.:

[0339] ;

[0340] Under the basic mapping relationship, C2 corresponds to the P2 doubt clarification path, C3 corresponds to the P3 behavioral reminder path, and C4 corresponds to the P4 continuation guarantee path. Therefore, the system includes P2, P3, and P4 in the candidate intervention path set. Subsequently, the system calculates the path priority scores for P2, P3, and P4 respectively. and in accordance with Sort the paths from highest to lowest. Paths listed earlier in the sorted list are output first; if a maximum number of output paths is set... Then the system output before The system selects the highest priority intervention path; if the minimum intervention combination rule is used, the system adds intervention paths in order of path priority until the obstacle coverage reaches the preset coverage threshold η.

[0341] When multiple obstacle subtypes coexist, the obstacle intensity, overall MIR score, expected utility of the pathway, and implementation burden can be comprehensively incorporated into the ranking process to prioritize the intervention pathways that are most relevant to the patient's current medication interruption risk, have higher expected effects, and are more feasible. This method supports both single-path priority output and multi-path combination output, and parameters can be adjusted based on expert consensus, historical follow-up data, and the management resources of different medical institutions.

[0342] The method for dynamically updating based on follow-up feedback data in step S7 includes:

[0343] The evaluation cycle is divided into two rounds, with the previous round designated as the preceding evaluation round and the current round as the current follow-up round. First, various feedback data generated during the current follow-up round are collected and used as the basic input for data updates in this round.

[0344] The collected feedback data undergoes standardization transformation: feedback indicating increased risk is uniformly converted to higher standardized risk values; feedback indicating decreased risk is uniformly converted to lower standardized risk values. If the feedback is continuous data or a scale scoring format, the theoretical highest and lowest scores for the corresponding content are used as reference standards to complete the numerical conversion within the 0 to 1 range; if the feedback is a binary option, it is directly assigned a value based on whether a risk has occurred.

[0345] After standardizing the feedback data, the risk sub-scores for the five dimensions of medication behavior, follow-up behavior, cognitive beliefs, supportive environment, and external environment are recalculated. These sub-scores are then combined with preset weights for each dimension to calculate the initial medication interruption risk score for this round. To avoid significant score deviations caused by fluctuations in single follow-up data, a smooth iteration method is used to determine the final medication interruption risk score for this round. This score is calculated by combining the risk scores from previous rounds with the initial score for this round, according to a preset update coefficient. The update coefficient can be set according to actual needs. When the update coefficient reaches its maximum value, the final score is determined entirely using the data from this round. When the update coefficient takes other values ​​within the range, a smooth update is achieved by considering both historical and current data.

[0346] Subsequently, the original obstacle intensity scores for the six types of medication discontinuation obstacles were recalculated. The score for each obstacle type was calculated by weighting the standardized risk values ​​of all its fields with corresponding weights. A smooth iterative approach was also used to update the final intensity score for each obstacle type, combining the obstacle scores from previous rounds, the original score for this round, and a specific update coefficient. Finally, based on the preset judgment thresholds for each type of obstacle and the updated obstacle intensity scores, the medication discontinuation obstacle type for the patient in this round was re-determined.

[0347] Based on the latest obstacle assessment results and obstacle intensity scores, and referring to the corresponding rules between obstacle classification and intervention paths, the activation conditions of each intervention path are recalculated to select candidate intervention paths that can be executed in this round.

[0348] Finally, the intervention tasks executed in this round are statistically analyzed, the task completion status is calculated, and the number of completed tasks is compared with the total number of tasks allocated by the system to obtain the task completion rate of the corresponding intervention path, thus completing the dynamic update of the entire process.

[0349] Specifically, with the first Round t represents the previous round of evaluation, and round t represents the follow-up feedback for this round. Feedback data is obtained in round t and is used as input data for the update in round t.

[0350] Feedback data includes:

[0351] Follow-up examination completion status: including whether blood lipids, blood sugar, blood pressure or related indicators have been re-examined;

[0352] Prescription renewal status: including whether the medication was renewed in a timely manner, and whether there were any instances where the medication ran out without being renewed;

[0353] Changes in missed doses: including the number of missed doses during this follow-up period and consecutive missed doses;

[0354] Changes in medication discontinuation: including whether the medication was discontinued on its own, the dosage was reduced on its own, or the medication was changed without authorization;

[0355] The completion status of intervention pathway tasks includes whether health education has been read, whether pharmacist consultation has been completed, whether medication check-in has been completed, whether family reminder tasks have been completed, and whether prescription renewal reminders from the community or the internet have been completed.

[0356] The feedback from medical staff includes nurse follow-up records, pharmacist consultation records, and doctor confirmation opinions.

[0357] The feedback data from this round was converted into standardized risk values. For feedback fields indicating increased risk, such as "still missed doses this round", "still not completed follow-up examination", "still not completed prescription renewal", and "discontinued medication again", they were converted into higher risk values. For feedback fields indicating decreased risk, such as "follow-up examination completed", "prescription renewal completed", "no more missed doses", and "pharmacist consultation completed", they were converted into lower risk values.

[0358] If a feedback field is a continuous variable or a scale score, it is converted to a risk value between 0 and 1 using a standardized formula. For risk fields, the following formula is used:

[0359] ;

[0360] For protected fields, use the following formula:

[0361] ;

[0362] in, For the first The standardized risk value of the q-th feedback field during the follow-up rounds;

[0363] For the first The actual value of the q-th feedback field during the follow-up round;

[0364] This is the theoretical minimum value for this feedback field;

[0365] This is the theoretical maximum value of the feedback field.

[0366] If the feedback field is a binary variable, then assign values ​​directly: risk not occurred = 0, risk occurred = 1; for example, in "whether medication was stopped on its own in this round", no = 0, yes = 1; "whether prescription was continued" is a protective field, completed = 0, not completed = 1.

[0367] According to the Feedback data is used to regenerate the medication behavior risk sub-score. Review of behavioral risk sub-scores Cognitive Belief Risk Sub-Score Support for environmental risk sub-scoring External environmental risk sub-score And calculate the original MIR score for this round:

[0368] ;

[0369] in, This is the original medication interruption risk score recalculated based on feedback data in round t;

[0370] For the risk sub-score of medication behavior in round t;

[0371] For the t-th round of review, the behavioral risk sub-score is used.

[0372] For the t-th round of cognitive belief risk sub-score;

[0373] For the t-th round of support, environmental risk sub-scores are provided;

[0374] For the external environment risk sub-score in round t;

[0375] , , , , The weighting coefficients of the corresponding sub-ratings, and the sum of the five is 1;

[0376] To avoid drastic changes in the MIR score due to fluctuations in a single feedback, a smooth update method is used to calculate the final MIR score in round t:

[0377] ;

[0378] in, The risk score for medication interruption after the t-th round of updates;

[0379] Assess the risk score for medication interruption in the previous round;

[0380] For the risk score of original medication interruption in round t;

[0381] These are the MIR update coefficients, with values ​​ranging from 0 to 1.

[0382] when When =1, it means that the MIR is recalculated using the feedback data from this round; when 0 < When the value is less than 1, it indicates a smooth update that combines the previous round of MIR scores with the feedback data from this round. It can be determined through expert consensus, historical follow-up data, or chronic disease management rules of medical institutions.

[0383] Recalculate the intensity scores for the six types of obstacles based on the feedback data from round t:

[0384] ;

[0385] in, The original obstacle intensity score for the j-th obstacle type in round t;

[0386] The standardized risk value of the k-th field in the j-th obstacle classification of round t;

[0387] The weight coefficient of the k-th field in the j-th obstacle classification;

[0388] The number of fields included in the j-th type of obstacle classification;

[0389] =1, 2, 3, 4, 5, 6, which correspond to C1 cognitive deficiency type, C2 medication concern type, C3 behavioral amnesia type, C4 difficulty in continuing prescription type, C5 insufficient support type, and C6 cost and information interference type, respectively.

[0390] The obstacle strength score is updated using a smoothing method:

[0391] ;

[0392] in, For the first The first round of updates Obstacle intensity score;

[0393] For the previous round Obstacle intensity score;

[0394] Score the original obstacle strength for this round;

[0395] This is the obstacle strength update coefficient, with a value range of 0-1;

[0396] According to the updated Reassess the first Wheel obstacle classification results:

[0397]

[0398] in, For the first Round Disorder classification results;

[0399] =1 indicates that the patient was at the first Such obstacles still exist in the wheel;

[0400] =0 indicates that the patient was in the first... The wheel did not meet the criteria for this type of obstacle.

[0401] For the first Threshold for classifying disorders; It can be determined through the Delphi expert inquiry method, the analytic hierarchy process, historical follow-up data, ROC curve analysis, or the chronic disease management parameter table of medical institutions.

[0402] According to the After the update of the obstacle classification results Obstacle Intensity Score Recalculate the trigger values ​​for each intervention path:

[0403] ;

[0404] in, For the first Round Trigger value for the intervention path;

[0405] For the first Classification of Disorders and the First Mapping relationships between intervention paths;

[0406] For the first Round Obstacle intensity score;

[0407] =1, 2, 3, 4, 5, 6, which correspond to the six intervention pathways from P1 to P6, respectively;

[0408] When >0, the first Class intervention path enters the first Round candidate path set; when =0,th Intervention-type paths are not included in the candidate path set;

[0409] No. The set of candidate paths for each round is represented as:

[0410] ;

[0411] in, Indicates the first A set of candidate intervention paths.

[0412] For the For the intervention paths that have already been executed, calculate the path task completion rate:

[0413] ;

[0414] in, For the first Round Task completion rate for intervention-type paths;

[0415] For the first Round The number of tasks completed in the intervention path;

[0416] For the first Round The total number of tasks assigned by the system in the intervention path.

[0417] Task completion rate can be used to judge the effectiveness of path execution. If a path has a high task completion rate and the corresponding MIR score or obstacle strength score decreases, the path can be downgraded in priority or removed from the candidate path in the next round; if the task completion rate is low or the corresponding risk has not been improved, the path can be retained, its content adjusted, or its execution method upgraded.

[0418] Preferably, the path priority can be recalculated by combining the updated MIR score, path trigger value, path expected utility value, and path implementation burden value:

[0419] ,

[0420] ;

[0421] in, For the first Round Priority scoring of intervention pathways;

[0422] For the first Round Trigger value for the intervention path;

[0423] For the first Updated medication interruption risk score;

[0424] For the first Round The expected utility value of intervention pathways;

[0425] For the first Round The implementation burden of intervention pathways.

[0426] according to Reorder the intervention paths from highest to lowest priority and output the priority paths for the next round. If the strength of an obstacle decreases after a certain original path is completed, its trigger value and priority will decrease accordingly; if an obstacle persists or worsens, its corresponding path will be retained or its priority will be increased in the next round.

[0427] According to the first The update results will be executed according to the following path output rules:

[0428] like Decrease, and corresponding A decrease indicates that the risk has improved after intervention, and the system can reduce the priority of this type of path or include it in routine follow-up.

[0429] like No decrease, and corresponding If the path does not decrease, it indicates that the current path has not improved sufficiently, and the system can retain the path and increase the follow-up intensity.

[0430] like A low score indicates insufficient completion of the path execution. The system can simplify the path tasks, add reminders, or transfer the follow-up to medical staff via manual follow-up.

[0431] If the newly added obstacle classification reaches the judgment threshold, the system will automatically generate a new intervention path;

[0432] If a certain type of obstacle classification falls below the judgment threshold for multiple consecutive rounds, the system can stop outputting the corresponding path or downgrade it to a regular reminder.

[0433] The method described above in this invention enables dynamic adjustment of MIR scores, medication interruption disorder classifications, and intervention pathways based on follow-up feedback data, so that intervention is no longer limited to a one-time assessment, but forms a continuous, updatable, and traceable closed-loop management process.

[0434] In this embodiment of the invention, the follow-up feedback dynamic update refers to obtaining feedback data generated by the patient, medical staff and system data sources after each round of intervention path execution, and re-incorporating the feedback data into the medication interruption risk score (MIR), obstacle intensity score and intervention path generation process, thereby realizing a closed-loop update of assessment, intervention, feedback, reassessment and re-intervention.

[0435] This invention also provides a system for identifying and generating intervention pathways for medication interruption risks in cardiovascular metabolic diseases, which is used to implement the method for identifying and generating intervention pathways for medication interruption risks in cardiovascular metabolic diseases as described above. The system structure includes a data acquisition module, a data standardization module, a feature generation module, an MIR risk scoring module, a medication interruption disorder classification module, an intervention pathway generation module, and a follow-up feedback update module.

[0436] Figure 1 The diagram illustrates the connection relationships between the patient terminal, medical staff terminal, system data source and data standardization module, feature generation module, MIR risk scoring module, medication interruption disorder classification module, intervention path generation module, and follow-up feedback update module in an embodiment of the present invention.

[0437] Figure 2 The following is an example of the sequential process of an embodiment of the present invention: S1 data acquisition, S2 data standardization, S3 feature generation, S4 MIR scoring, S5 obstacle classification, S6 path generation, and S7 feedback update.

[0438] Figure 3 The paper illustrates the correspondence between the following types of patients in this invention: C1 (insufficient cognition), C2 (medication concerns), C3 (behavioral amnesia), C4 (difficulty in prescription renewal), C5 (insufficient support), and C6 (cost and information interference), and the following paths: P1 (goal cognition path), P2 (doubt clarification path), P3 (behavioral reminder path), P4 (prescription renewal guarantee path), P5 (family collaboration path), and P6 (accessibility and information correction path). It also displays the minimum intervention priority ranking formula.

[0439] The present invention can also be extended to the following alternatives, as follows:

[0440] 1. Alternative data sources. Data can come from patient-side mobile applications, WeChat mini-programs, hospital HIS systems, electronic medical record systems, laboratory systems, prescription systems, internet hospitals, community health service platforms, or wearable devices, or a combination of the above sources.

[0441] 2. Alternative solutions for MIR scoring. MIR scoring can be implemented using expert rule scoring, weighted summation scoring, logistic regression models, decision tree models, random forest models, gradient boosting models, neural network models, or other machine learning models.

[0442] 3. Alternative solutions for obstacle classification. Obstacle classification can output a single primary classification or a combination of multiple classifications; it can be identified using rule-based thresholding, cluster analysis, classification models, or multi-label classification models.

[0443] 4. Alternative Intervention Pathways. Intervention pathways can be generated automatically or recommended by the system and then confirmed by medical staff. The pathway content can be output in the form of text, task lists, SMS, App reminders, telephone follow-up tasks, pharmacist consultation tasks, family reminder tasks, etc.

[0444] 5. Alternative to the minimum intervention priority ranking. The parameters Bj, MIR, Uj, and Mj in the priority ranking formula can be adjusted based on different institutions, diseases, and drug types. Parameters such as patient preferences, healthcare resources, and service accessibility can also be incorporated.

[0445] 6. Feedback Update Alternatives. Feedback updates can be performed at fixed time intervals, such as weekly, monthly, or quarterly; or they can be performed when a triggering event occurs, such as consecutive missed doses, medication running out, missed follow-up appointments, or when the patient proactively reports discomfort.

[0446] 7. Applicable Disease Alternatives. This invention can be used for the management of lipid-lowering drugs in patients with early-onset type 2 diabetes, and also for the long-term medication management of patients with diabetes, hypertension, dyslipidemia, obesity, coronary heart disease, high-risk stroke, or other cardiovascular and metabolic diseases.

[0447] 8. Alternative System Deployment Solutions. The system can be deployed on hospital servers, cloud servers, internet healthcare platforms, community chronic disease management platforms, or in-hospital nursing follow-up platforms.

[0448] This invention also provides a computer device. Figure 4 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present invention; see the attached drawing. Figure 4 As shown, the computer device includes: an input device 23, an output device 24, a memory 22, and a processor 21; the memory 22 is used to store one or more programs; when the one or more programs are executed by the one or more processors 21, the one or more processors 21 implement the cardiovascular metabolic disease medication interruption risk identification and intervention path generation method provided in the above embodiments; wherein the input device 23, the output device 24, the memory 22, and the processor 21 can be connected via a bus or other means. Figure 4 Taking the example of a connection between China and Israel via a bus.

[0449] The memory 22, as a read / write storage medium for a computing device, can be used to store software programs and computer-executable programs, such as the program instructions corresponding to the cardiovascular metabolic disease medication interruption risk identification and intervention path generation method described in this embodiment of the invention. The memory 22 may mainly include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a function; the data storage area may store data created based on the use of the device. Furthermore, the memory 22 may include high-speed random access memory and non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 22 may further include memory remotely located relative to the processor 21, and these remote memories can be connected to the device via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0450] Input device 23 can be used to receive input digital or character information, and generate key signal inputs related to user settings and function control of the device; output device 24 may include display devices such as a display screen.

[0451] The processor 21 executes various functional applications and data processing of the device by running software programs, instructions and modules stored in the memory 22, thereby realizing the above-mentioned method for identifying the risk of medication interruption in cardiovascular metabolic diseases and generating intervention pathways.

[0452] The computer equipment provided above can be used to execute the cardiovascular metabolic disease medication interruption risk identification and intervention path generation method provided in the above embodiments, and has corresponding functions and beneficial effects.

[0453] This invention also provides a storage medium containing computer-executable instructions, which, when executed by a computer processor, are used to perform the cardiovascular metabolic disease medication interruption risk identification and intervention path generation method provided in the above embodiments. The storage medium can be any type of memory device or storage device, including: mounting media such as CD-ROM, floppy disk, or magnetic tape; computer system memory or random access memory such as DRAM, DDRRAM, SRAM, EDORAM, Rambus RAM, etc.; non-volatile memory such as flash memory, magnetic media (e.g., hard disk or optical storage); registers or other similar types of memory elements; the storage medium may also include other types of memory or combinations thereof; furthermore, the storage medium may reside in a first computer system in which the program is executed, or it may reside in a different second computer system connected to the first computer system via a network (such as the Internet); the second computer system can provide program instructions to the first computer for execution. The storage medium includes two or more storage media that may reside in different locations (e.g., in different computer systems connected via a network). The storage medium may store program instructions (e.g., specifically implemented as a computer program) executable by one or more processors.

[0454] Of course, the computer-executable instructions provided in the embodiments of the present invention are not limited to the method for identifying and generating intervention pathways for drug interruption in cardiovascular metabolic diseases as described in the above embodiments, but can also perform related operations in the method for identifying and generating intervention pathways for drug interruption in cardiovascular metabolic diseases provided in any embodiment of the present invention.

[0455] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

[0456] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for identifying the risk of medication discontinuation in cardiovascular metabolic diseases and generating an intervention path, characterized by, Includes the following steps: S1. Obtain multi-source data of patients with cardiovascular and metabolic diseases, including disease information, medication information, medication behavior information, follow-up behavior information, cognitive belief information, support environment information, and external environment information; S2. Standardize the multi-source data, including desensitization, encoding, scoring standardization, time window standardization, and anomaly and missing data marking; S3. Generate medication behavior characteristics, follow-up behavior characteristics, cognitive belief characteristics, support environment characteristics, and external environment characteristics based on standardized data; S4. The medication interruption risk score (MIR) is calculated by combining the generated features. The medication interruption risk score (MIR) is calculated by weighting the risk assessment results of five dimensions: medication behavior, follow-up behavior, cognitive beliefs, support environment, and external environment. The individual assessment results of the above five dimensions have been standardized and the value range is uniformly set to 0 to 1. The larger the value of a single item, the higher the risk of medication interruption for the corresponding dimension. Each dimension is assigned a corresponding weight coefficient, and the sum of all weight coefficients is 1; the final medication interruption risk score MIR ranges from 0 to 100. The higher the MIR score, the higher the overall risk of the patient experiencing medication interruption. S5. Classify patients with medication interruption disorder, including C1 cognitive deficiency, C2 medication anxiety, C3 behavioral amnesia, C4 difficulty in continuing medication, C5 lack of support, and C6 cost and information interference. S6. Generate corresponding intervention paths based on the obstacle classification results; S7. Prioritize multiple intervention pathways by minimum intervention level; dynamically update MIR scores, impairment classifications, and intervention pathways based on follow-up feedback data.

2. The cardiovascular metabolic disease medication discontinuation risk identification and intervention pathway generation method according to claim 1, characterized by, In step S4, the sub-scores of risk for each dimension are calculated as follows: The medication behavior risk sub-score is calculated based on the medication coverage during the patient's observation period. Medication coverage is the proportion of days with medication coverage to the total number of days in the observation period. The medication behavior risk sub-score increases as the medication coverage ratio decreases. When the medication coverage ratio is lower than a preset threshold, it is judged as insufficient medication coverage and is included in the medication interruption risk score. The risk sub-score for follow-up behavior is calculated based on the patient's actual follow-up interval and the upper limit of follow-up time recommended by guidelines or consensus. The risk sub-score for review behavior ranges from 0 to 1. When the actual review interval does not exceed the upper limit of the recommended time, the risk sub-score for review behavior is 0. When the actual review interval exceeds the upper limit of the recommended time, the risk sub-score for review behavior increases as the proportion of the time exceeding the upper limit of the recommended time increases, but does not exceed 1. Standardized calculations were performed on the corresponding dimensions of the questionnaire, which had undergone reliability and validity testing, to obtain sub-scores for cognitive belief risk, supporting environment risk, and external environment risk. For questionnaire items with protective statements, the higher the actual score, the lower the risk, and these scores were converted into risk values ​​that were positively correlated with the degree of risk. For questionnaire items with risky statements, the higher the actual score, the higher the risk, and these scores were directly converted into risk values ​​that were positively correlated with the degree of risk. During the conversion process, the theoretical minimum and theoretical maximum scores of the corresponding dimensions are used as benchmarks to perform a linear mapping on the actual scores to obtain a standardized risk value. Then, based on this risk value, the corresponding cognitive belief risk sub-score, supporting environment risk sub-score, or external environment risk sub-score is calculated.

3. The cardiovascular metabolic disease medication discontinuation risk identification and intervention pathway generation method according to claim 1, characterized by, Step S5, the method for classifying patients as having medication discontinuation disorders, includes: S51. For each type of obstacle, extract the original fields related to that obstacle and convert them into standardized risk values ​​between 0 and 1. For fields describing risk, the higher the actual risk level, the larger the standardized risk value after conversion; for fields describing protection, the higher the actual protection level, the smaller the standardized risk value after conversion. The conversion process uses the theoretical minimum and theoretical maximum scores of the corresponding field as benchmarks and performs a linear mapping to the actual scores to obtain standardized risk values. If the field is a binary variable, it is directly assigned a value of 0 or 1 based on whether there is risk. If the field is a ternary or multi-category variable, it is converted into a value between 0 and 1 according to the risk level from low to high. S52. Calculate the obstacle intensity score for each of the six obstacle types, C1 to C6. The obstacle intensity score is calculated by weighted average of the standardized risk values ​​of each relevant field under the corresponding type. Each field is configured with a corresponding weight coefficient, and the weight coefficient of the field is positively correlated with the degree of influence of the field on the obstacle type. The value range of the six obstacle intensity scores is 0 to 1. The higher the score, the higher the intensity of the medication interruption obstacle of that type. C1 is the cognitive deficiency type, C2 is the medication concern type, C3 is the behavioral amnesia type, C4 is the prescription refill difficulty type, C5 is the lack of support type, and C6 is the cost and information interference type.

4. The cardiovascular metabolic disease medication discontinuation risk identification and intervention pathway generation method according to claim 3, characterized by, In step S5, after calculating the intensity scores of various obstacles, the identification and judgment of medication interruption obstacles are completed in combination with the preset judgment threshold. Each type of obstacle corresponds to an independent judgment threshold, which can be determined by Delphi expert inquiry method, analytic hierarchy process, historical follow-up data, ROC curve analysis or chronic disease management parameter table of medical institutions. The calculated obstacle intensity score is compared with the corresponding preset threshold. When the obstacle intensity score of a certain type is greater than or equal to the corresponding threshold, the patient is determined to have that type of medication interruption obstacle; when the obstacle intensity score is less than the corresponding threshold, the patient is determined not to have that type of medication interruption obstacle. When outputting the classification results, different output formats can be selected according to actual needs: if it is only necessary to determine the patient's main disorder type, the category with the highest disorder intensity score is selected from the six types of disorders and it is determined as the patient's main medication interruption disorder type; if the intensity scores of two or more disorders reach their respective judgment thresholds, the corresponding combinations of disorders are output simultaneously, indicating that the patient has multiple medication interruption disorders at the same time.

5. The cardiovascular metabolic disease medication discontinuation risk identification and intervention pathway generation method according to claim 1, characterized by, The intervention pathway in step S6 includes: P1 goal cognition pathway, P2 doubt clarification pathway, P3 behavior reminder pathway, P4 prescription renewal guarantee pathway, P5 family collaboration pathway, and P6 accessibility and information correction pathway. Methods for generating corresponding intervention paths based on obstacle classification results include: S61. Data reception: Obtain the results of the six types of medication interruption obstacles obtained in the aforementioned steps, as well as the obstacle intensity score corresponding to each type of obstacle. The obstacle intensity score ranges from 0 to 1, and the higher the score, the more prominent the corresponding obstacle is. S62. Pre-establish the corresponding association rules between the disorder classification and the intervention path, clarify the matching relationship between the six types of medication interruption disorders and the six types of intervention paths, and adopt a one-to-one correspondence basic association form. Determine whether a certain disorder can trigger the corresponding intervention path based on the association rules. S63. Combining the obstacle determination results, obstacle intensity scores, and preset association rules, comprehensively calculate the triggering degree of each type of intervention path; if the triggering degree of a certain intervention path is greater than zero, then the generation of that type of intervention path is initiated; if the triggering degree is zero, then the corresponding intervention path is not generated; in the one-to-one correspondence association mode, when it is determined that a patient has a certain type of medication interruption obstacle, the intensity score of that type of obstacle is the triggering basis for the corresponding intervention path, and a matching intervention path is generated accordingly. S64. After determining the intervention path to be generated, call the corresponding intervention task package according to the path type.

6. The cardiovascular metabolic disease medication discontinuation risk identification and intervention pathway generation method according to claim 5, characterized by, Step S64, which involves calling the corresponding intervention task package based on the path type, includes: P1 Goal Cognition Path: Corresponding to C1 Cognitive Deficiency Type, the task package includes explanation of long-term medication goals, explanation of disease indicator target values, education on the necessity of continuous management, and explanation of the purpose of follow-up examination; P2 Concern Clarification Path: Corresponding to C2 Drug Concern Type, the task package includes identification of common adverse drug reactions, consultation reminders when discomfort occurs, and pharmacist or medical staff consultation tasks. P3 Behavioral Reminder Path: Corresponds to C3 Behavioral Forgetting Type, the task package includes fixed medication time reminders, medication check-in, and triggering medical staff follow-up for consecutive missed doses; P4 Renewal Guarantee Pathway: Corresponding to the C4 Renewal Difficulty Type, the task package includes advance renewal reminders, community renewal prompts, online renewal prompts, and reminders for review and renewal coordination; P5 Family Collaboration Path: Corresponding to the C5 insufficient support type, the task package includes family education, joint reminders for family members, reminders for accompanying family members to follow-up visits, and family members' participation in follow-up tasks. P6 Accessibility and Information Correction Path: Corresponding to C6 Cost and Information Interference Type, the task package includes cost burden assessment, medical insurance or reimbursement reminders, drug accessibility reminders, correction of network misinformation, and risk warnings of health product substitution. 7.The cardiovascular metabolic disease medication discontinuation risk identification and intervention pathway generation method of claim 1, wherein, The method for prioritizing the minimum intervention paths in step S7 includes: S71. Screening candidate intervention paths: Based on the obstacle classification results, obstacle intensity scores, and preset obstacle-intervention path association rules, determine whether each intervention path meets the activation conditions; intervention paths that meet the activation conditions are included in the candidate path set, while paths that do not meet the activation conditions are directly excluded. S72. Calculate the priority score for each candidate path: For each intervention path in the candidate set, calculate the priority score by considering multiple dimensions. The scoring reference dimensions include the degree of path triggering, the overall risk of medication interruption for patients, the expected improvement effect of the intervention path, and the actual difficulty of path execution. The overall risk of medication interruption is normalized according to the standard range. The lower the difficulty of path execution, the higher the corresponding score ratio. S73. Determine the expected utility value of the intervention path: The expected utility value of the intervention path is used to measure the estimated effect of the path in improving medication continuity, reducing the risk of medication interruption, or eliminating barriers to medication use. The utility value ranges from 0 to 1, with higher values ​​indicating more ideal expected improvement. The expected utility value can be calculated based on historical follow-up data, mainly comparing the change in the overall risk of medication interruption before and after the implementation of the same type of intervention path. If the path targets a single medication barrier, the change in the intensity of the corresponding barrier before and after the intervention is compared to determine the actual utility level of the path. S74. Determine the implementation burden of the intervention path: The implementation burden of the intervention path is used to characterize the difficulty of implementing the path. The value range is from 0 to 1. The higher the value, the greater the implementation pressure. The implementation burden is comprehensively evaluated by combining four factors: the difficulty of operation for patients, the difficulty of cooperation for family members, the workload of medical staff, and the economic costs incurred. Each factor is assigned a corresponding weight, and the final implementation burden result is obtained by integrating and calculating according to the weight ratio. S75. Complete the path sorting and result output: Based on the calculated priority score, the candidate intervention paths are sorted in descending order of score; in actual application, the upper limit of the number of paths that can be executed in a single follow-up can be preset according to the task capacity of the medical institution's follow-up management, and finally, the top-ranked intervention paths are selected as the execution plan for this time and output.

8. The method for identifying the risk of medication interruption in cardiovascular metabolic diseases and generating intervention pathways according to claim 1, characterized in that, The method for dynamically updating based on follow-up feedback data in step S7 includes: The evaluation cycle is divided into two rounds, with the previous round being designated as the preceding evaluation round and the current round as the current follow-up round. First, various feedback data generated during the current follow-up round are collected and used as the basic input for data updates in this round. The collected feedback data undergoes standardized transformation: feedback indicating increased risk is uniformly converted to higher standardized risk values; feedback indicating reduced risk is uniformly converted to lower standardized risk values; if the feedback is continuous data or a scale scoring format, the theoretical highest and lowest scores for the corresponding content are used as reference standards to complete the numerical conversion within the 0 to 1 range; if the feedback is a binary option, it is directly differentiated according to whether a risk has occurred. After standardizing the feedback data, the risk sub-scores for the five dimensions of medication behavior, follow-up behavior, cognitive beliefs, supportive environment, and external environment are recalculated. Combined with the preset weights for each dimension, the initial medication interruption risk score for this round is calculated. To avoid significant score deviations caused by fluctuations in single follow-up data, a smooth iteration method is used to determine the final medication interruption risk score for this round. This score is calculated by combining the risk scores from previous rounds with the initial score for this round, according to a preset update coefficient. The update coefficient can be set according to actual needs. When the update coefficient is at its maximum value, the final score is determined entirely using the data from this round. When the update coefficient is at other values ​​within the range, a smooth update is achieved by considering both historical and current data. The original obstacle intensity scores for the six types of medication interruption obstacles were recalculated. The score for each obstacle was calculated by weighting the standardized risk values ​​of all fields in the category with the corresponding weights. The final intensity score for each obstacle was updated using a smooth iterative method, combining the obstacle scores from previous rounds, the original score for this round, and the specific update coefficients to complete the calculation. Then, based on the preset judgment thresholds for each type of obstacle and the updated obstacle intensity scores, the medication interruption obstacle type for the patient in this round was re-determined. Based on the latest obstacle assessment results and obstacle intensity scores, and referring to the corresponding rules between obstacle classification and intervention paths, the activation conditions of each intervention path are recalculated to select candidate intervention paths that can be executed in this round. The system compiles statistics on the intervention tasks executed in this round, calculates the task completion status, compares the number of completed tasks with the total number of tasks allocated by the system, obtains the task completion rate for the corresponding intervention path, and dynamically updates the entire process.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps of the method for identifying and generating intervention pathways for medication interruption in cardiovascular metabolic diseases as described in any one of claims 1-8.

10. A computer device, the computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the method for identifying and generating intervention pathways for medication interruption in cardiovascular metabolic diseases as described in any one of claims 1-8.

Images