An intelligent intervention system for personalized lifestyle of the elderly comorbidity population

Abstract

The present application relates to the technical field of health data processing, and discloses a kind of personalized lifestyle intelligent intervention system for old comorbidity population.The present application system receives structured disease diagnosis information, medication information, old comprehensive assessment information and multi-source candidate lifestyle intervention measures, writes candidate measures into candidate intervention record, and converts the above information into constraint set;According to intervention object and constraint category composition rule index key, call target conflict rule set, and match intervention parameter, structured field value and trigger condition field.Result hit is written into reservation mark, execution condition, execution parameter, deletion mark, alternative measure call mark or review mark by disposal execution module;Alternative candidate intervention record and closed loop update form adjusted candidate intervention record, which reenters rule index, conflict detection and disposal execution process before writing intervention package.

Description

Technical Field

[0001] This invention relates to the field of health data processing technology, and in particular to a personalized lifestyle intelligent intervention system for elderly people with comorbidities. Background Technology

[0002] Elderly individuals with comorbidities typically present with multiple chronic diseases, medication records, declining functional status, fall risk, cognitive changes, nutritional risks, and insufficient care resources. The generation and delivery of lifestyle interventions in this population should not rely solely on single disease labels or general user profiles to directly formulate dietary, exercise, sleep, psychological, social, or medication adherence recommendations. Instead, a unified approach is needed to process the relationships between candidate interventions and disease diagnosis, medication information, and comprehensive geriatric assessment information before outputting them. If different intervention modules generate candidate measures separately without a unified mechanism for recording, indexing, matching trigger conditions, and writing treatments, it can easily lead to the same user receiving multiple inconsistent candidate interventions simultaneously, and the closed-loop update phase bypassing existing constraint matching and treatment controls.

[0003] Existing systems related to health management for chronic diseases in the elderly typically revolve around health records, risk assessment, physician management, individualized intervention, self-management, and follow-up feedback. For example, patent CN112164455B discloses an interactive health management system and its management method for chronic diseases in the elderly. This system includes a physician decision-making system, an individualized intervention system, and a self-management system. It can collect relevant information about elderly patients with chronic diseases, generate health assessment reports, and provide intervention guidance on diet, exercise, and psychology. This type of system focuses on the health management process and individualized intervention services for chronic diseases in the elderly. However, it lacks specific solutions for how to uniformly incorporate multiple candidate lifestyle interventions (such as diet, exercise, sleep, psychology, social interaction, and medication adherence) into candidate intervention records, and how to invoke and handle conflict rules based on the intervention recipients and user constraint categories of candidate interventions.

[0004] Existing solutions related to lifestyle medicine management for chronic diseases typically output a comprehensive prescription encompassing diet, nutrition, exercise, and psychological interventions. For example, patent CN110491526A discloses a multi-dimensional intervention-based lifestyle medicine management system for chronic diseases. This system conducts a comprehensive screening and assessment based on the patient's health status, dietary status, nutritional status, physical fitness status, and psychological status, and then issues a personalized lifestyle prescription. While these solutions focus on the generation and management of multi-dimensional lifestyle prescriptions, with a key emphasis on forming intervention recommendations based on assessment results, they lack a complete information processing workflow. This workflow is inadequate for tasks such as indexing and matching each candidate lifestyle intervention with disease constraints, medication constraints, and comprehensive geriatric assessment constraints using field-based rules before prescription output, and for merging treatment parameters, writing and deleting verification markers when multiple rules are simultaneously applied.

[0005] Existing AI-based chronic disease follow-up and monitoring solutions typically employ knowledge graphs, risk prediction models, user profiles, and follow-up strategy generation mechanisms. For example, patent CN113643813A discloses an AI-based method, device, and computer equipment for chronic disease follow-up and monitoring. This solution can collect chronic disease pathological data, extract pathological features, construct risk prediction models and user profiles, and generate follow-up monitoring strategies, medication reminders, dietary and nutritional plans, and dietary and exercise health recommendations. Such solutions focus on chronic disease risk prediction, user profiling, and follow-up strategy updates, emphasizing intelligent management of user status and follow-up strategies. However, for intervention adjustment requests generated after follow-up feedback, there is no mechanism to re-trigger the rule index, conflict detection, and handling execution process before updating the lifestyle intervention package; instead, a re-verification mechanism combining candidate intervention records is lacking.

[0006] Existing exercise prescriptions or health intervention programs for the elderly typically focus on exercise risk assessment, exercise prescription recommendation, movement detection, or exercise process monitoring. For example, patent CN114373549B discloses an adaptive exercise prescription health intervention method and system for the elderly. This program assesses exercise risk for the elderly, determines trial exercise prescriptions, and matches target exercise prescriptions based on exercise monitoring videos and movement characteristics. While this type of program primarily addresses the adaptive adjustment of the exercise prescription itself, it lacks a systematic approach to how to integrate candidate interventions from different sources into a unified candidate intervention set and manage them under the same constraint matching process when exercise candidate interventions coexist with candidate interventions related to diet, sleep, psychology, social interaction, and medication adherence.

[0007] Existing elderly health monitoring or early warning systems typically revolve around vital sign collection, environmental monitoring, sudden illness identification, abnormal warnings, and the transmission of assistance requests. For example, patent CN117133464B discloses an intelligent health monitoring system and method for the elderly. This system collects environmental conditions and vital sign data from the elderly's activity areas, combines this data with a sudden illness knowledge graph to determine suspected illnesses, and matches emergency plans and assistance requests. Such systems primarily process monitoring data, abnormal identification results, and early warning information. They do not focus on the process of using abnormal event records as re-verification input before adjusting lifestyle intervention packages, or on the process of generating candidate measure deactivation requests, execution parameter downgrade requests, alternative candidate call requests, or review mark requests based on abnormal event records, followed by re-matching the adjusted candidate intervention measures according to conflict rules.

[0008] Furthermore, existing programs related to frailty, cognitive function, or comprehensive assessment are typically used to identify frailty status, cognitive risk, or changes in functional status in older adults. While these programs can provide comprehensive assessment results, their primary function is assessment or early warning. Information such as frailty status, fall risk, cognitive status, nutritional risk, daily living abilities, and care resources is usually not further transformed into a set of constraints that can be used for rule indexing and trigger condition matching, nor is a unified data processing chain established with the intervention targets, frequency, duration, intensity, and intervention parameters of candidate lifestyle interventions.

[0009] In summary, the existing technology has at least the following shortcomings: First, existing chronic disease management and lifestyle prescription systems for the elderly focus primarily on health assessment, risk prediction, dietary, exercise, and psychological recommendations, and follow-up management. They typically do not treat candidate lifestyle interventions as candidate intervention records for unified processing and lack a process for recording, indexing, and field-based treatment of candidate interventions for elderly comorbidity scenarios.

[0010] Second, while existing multidimensional lifestyle intervention programs can generate comprehensive recommendations on diet, exercise, and psychology, they typically do not match the intervention parameters of candidate intervention measures with the constraint set formed by disease diagnosis information, medication information, and comprehensive geriatric assessment information before outputting the recommendations. They also do not perform actions such as retaining marker writing, executing condition writing, adjusting execution parameters, deleting marker writing, calling alternative measures, or reviewing marker writing based on the hit results.

[0011] Third, existing intelligent follow-up and feedback adjustment schemes typically update recommendations or follow-up strategies directly after collecting feedback, lacking a re-verification mechanism to re-trigger rule indexing, conflict detection, and disposal execution before updating the lifestyle intervention package. This can easily lead to adjusted candidate intervention measures bypassing the original constraint matching and disposal control processes.

[0012] Fourth, existing technologies for elderly health monitoring, exercise prescription, and comprehensive elderly assessment mostly exist as separate functional modules, and have not formed a unified information processing system that takes candidate lifestyle interventions as the processing object, constraint sets as the matching basis, conflict rule fields as the calling object, and the writing of treatment parameters as the output method.

[0013] Therefore, there is a need for a personalized lifestyle intelligent intervention system for elderly people with comorbidities. This system should unify the processing of structured disease diagnosis information, medication information, comprehensive geriatric assessment information, and multi-source candidate lifestyle intervention measures, so that candidate intervention measures can enter a consistent constraint matching and treatment control process before the lifestyle intervention package is generated and updated. Summary of the Invention

[0014] Given that existing technologies typically generate lifestyle interventions for elderly people with comorbidities separately from different modules such as diet, exercise, sleep, psychology, social interaction, and medication adherence, and that there is a lack of a unified process for recording candidate interventions, matching constraint sets, indexing conflict rules, matching trigger conditions, and writing treatment measures before these interventions are written into the lifestyle intervention package, and that the closed-loop adjustment phase can easily bypass the original constraint matching and treatment control issues, a personalized intelligent lifestyle intervention system for elderly people with comorbidities is proposed.

[0015] Its purpose is to unify the information processing of structured disease diagnosis information, medication information, comprehensive geriatric assessment information, and multi-source candidate lifestyle interventions, convert candidate lifestyle interventions into a candidate intervention set, and before the lifestyle intervention package is generated and updated, to perform retention mark writing, execution condition writing, execution parameter adjustment, deletion mark writing, alternative measure invocation or review mark writing on candidate interventions through rule indexing, trigger condition matching, treatment action field reading and candidate intervention record writing, so that candidate interventions enter a consistent constraint matching and treatment control process.

[0016] The technical solution of this invention is a personalized lifestyle intelligent intervention system for elderly people with comorbidities, including a data receiving module, a constraint set generation module, a candidate intervention set generation module, a rule indexing module, a conflict detection module, a treatment execution module, an intervention package generation module, and a closed-loop re-verification module.

[0017] The data receiving module receives structured disease diagnosis information, medication information, comprehensive geriatric assessment information, and candidate lifestyle interventions; the constraint set generation module converts the disease diagnosis information, medication information, and comprehensive geriatric assessment information into constraint sets.

[0018] The candidate intervention set generation module writes each candidate intervention measure into a candidate intervention record, and multiple candidate intervention records form a candidate intervention set; the rule indexing module forms a rule index key based on the intervention object in the candidate intervention record and the constraint category in the constraint set, and calls the target conflict rule set from the conflict rule base according to the rule index key.

[0019] The conflict detection module matches the intervention parameters in the candidate intervention record and the structured field values ​​in the constraint set with the trigger condition fields of each target conflict rule in the target conflict rule set to generate the target conflict rule hit result; when the target conflict rule hit result is empty, the disposal execution module performs a retention mark writing on the candidate intervention record.

[0020] When the result of the target conflict rule hit is not empty, read the action field, parameter field and priority field of the target conflict rule corresponding to the result of the target conflict rule hit, and write the execution conditions, adjust the execution parameters, write the candidate measure deletion mark and write the alternative measure call or review mark according to the priority field.

[0021] The intervention package generation module writes candidate intervention records that have been processed by the treatment execution module and have not been marked for deletion into the lifestyle intervention package.

[0022] The closed-loop re-verification module receives compliance data, physiological monitoring feedback data, and abnormal event records corresponding to the lifestyle intervention package, generates an intervention package adjustment request, records the adjusted candidate intervention measures involved in the intervention package adjustment request as adjusted candidate intervention records, and re-sends the adjusted candidate intervention records into the rule index module, conflict detection module, and disposal execution module for processing before the lifestyle intervention package is updated.

[0023] Furthermore, the constraint set includes disease constraint subsets, medication constraint subsets, and comprehensive geriatric assessment constraint subsets.

[0024] The constraint set generation module generates disease constraint subsets based on disease diagnosis information, medication constraint subsets based on medication information, and comprehensive geriatric assessment constraint subsets based on comprehensive geriatric assessment information.

[0025] The constraint subset of the comprehensive assessment for the elderly includes functional status constraint records transformed from frailty status field, fall risk field, cognitive status field, nutritional risk field, daily living ability field, and care resource field.

[0026] Furthermore, the constraint set generation module converts the frailty state field into frailty constraints, the fall risk field into fall constraints, the cognitive state field into execution complexity constraints, the nutritional risk field into dietary execution constraints, the daily living ability field into task execution constraints, and the care resource field into care assistance constraints, and writes the conversion results into the constraint subset of the comprehensive geriatric assessment.

[0027] Furthermore, the candidate intervention set generation module writes each candidate lifestyle intervention into a candidate intervention record.

[0028] The candidate intervention record includes candidate source, intervention category, intervention target, execution frequency, execution duration, execution intensity, execution time, target parameters, and intervention parameters.

[0029] Intervention categories include dietary intervention, exercise intervention, sleep intervention, psychological intervention, social intervention, and medication adherence intervention.

[0030] Furthermore, the conflict rule library called by the rule index module stores multiple conflict rules.

[0031] Each conflict rule includes a rule identifier, a rule version identifier, a rule source identifier, a constraint category field, a constraint object field, an intervention object field, a trigger condition field, a disposal action field, a disposal parameter field, and a disposal priority field.

[0032] Rule identifiers are used to distinguish different conflicting rules in the conflict rule base, rule version identifiers are used to record the version of the conflicting rule, and rule source identifiers are used to record the source of the conflicting rule.

[0033] The constraint category field stores constraint category values ​​that match the constraint categories in the constraint set.

[0034] The constraint object field stores the constraint object value that matches the structured field value in the constraint set.

[0035] The intervention object field stores the intervention object value that matches the intervention object in the candidate intervention record.

[0036] The trigger condition field stores field matching conditions, field missing conditions, tag hit conditions, threshold comparison conditions, or intervention parameter out-of-bounds conditions.

[0037] Actions include writing conditions to the disposal action field, parameter adjustment actions, writing candidate measure deletion markers, calling alternative measures, or writing review markers.

[0038] The disposal parameter field stores the write parameters, adjustment parameters, replacement matching parameters, or review field parameters corresponding to the disposal action field.

[0039] The priority field stores the execution sorting parameters when multiple conflicting rules for a target are hit simultaneously.

[0040] Furthermore, the management terminal receives rule source information, applicable objects, applicable fields, comparison operators, threshold fields, enumeration boundaries, trigger condition fields, handling action fields, handling parameter fields, and handling priority fields, and generates conflict rule records to be added to the database based on preset field templates.

[0041] The management side performs field integrity checks and field type checks on the conflict rule records to be entered into the database.

[0042] Field integrity verification includes checking whether the rule identifier, rule version identifier, rule source identifier, constraint category field, constraint object field, intervention object field, trigger condition field, action field, and action parameter field are empty.

[0043] Field type validation includes validating whether the threshold field is a numeric field, whether the enumeration boundary belongs to the preset enumeration range, whether the comparison operator belongs to the preset comparison operator set, whether the trigger condition field can be parsed as a field matching condition, field missing condition, tag hit condition, threshold comparison condition or intervention parameter out-of-bounds condition, and whether the action field belongs to one of the following: condition writing action, parameter adjustment action, candidate measure deletion tag writing action, alternative measure invocation action or review tag writing action.

[0044] After a conflict rule record to be added to the database passes the field integrity check and field type check, the management terminal writes the conflict rule record to the conflict rule database. If a conflict rule record to be added to the database fails the field integrity check or field type check, the management terminal marks the conflict rule record as having an error in the rule configuration and prevents the conflict rule record from entering the callable rule set of the rule index module.

[0045] Furthermore, the rule indexing module establishes rule index keys based on the constraint category field and the intervention object field; the rule indexing module reads the intervention object from the candidate intervention record, reads the constraint category with written constraint value from the constraint set, and uses the intervention object and the constraint category with written constraint value to form the rule index key, and reads the target conflict rule set from the conflict rule base.

[0046] Furthermore, the conflict detection module reads the constraint object field and trigger condition field of each target conflict rule in the target conflict rule set.

[0047] The conflict detection module reads the structured field values ​​from the constraint set and the intervention parameters from the candidate intervention records.

[0048] The conflict detection module performs equality matching on structured field values ​​and field matching conditions, matches empty fields with field missing conditions, matches structured field values ​​with tag hit conditions, compares structured field values ​​with threshold comparison conditions, and compares intervention parameters with upper and lower limits of intervention parameter out-of-bounds conditions.

[0049] The conflict detection module generates a target conflict rule hit result when the structured field value corresponding to the constraint object field matches the trigger condition field, or when the intervention parameter in the candidate intervention record matches the trigger condition field.

[0050] Furthermore, when the same candidate intervention record matches multiple target conflict rules, the handling execution module sorts the multiple target conflict rules according to the handling priority field.

[0051] The disposal execution module writes execution conditions, adjusts execution parameters, writes candidate measure deletion markers, and writes alternative measure invocation or review markers for candidate intervention records according to the disposal action field and disposal parameter field.

[0052] The treatment execution module will merge treatment parameters that can be written to the same candidate intervention record at the same time and write them into the candidate intervention record.

[0053] When the action execution module writes the same parameter position in multiple action fields and the values ​​are inconsistent, or when the execution objects of multiple action fields are mutually exclusive, it writes a review mark to the candidate intervention record.

[0054] When the disposal action field contains the action to write a candidate intervention deletion flag, the disposal execution module writes a deletion flag to the candidate intervention record to prevent it from being included in the lifestyle intervention package.

[0055] Furthermore, after the candidate intervention record is marked with a deletion tag, the treatment execution module reads the intervention category and intervention target of the candidate intervention record with the deletion tag.

[0056] The disposal execution module retrieves alternative candidate measures from the alternative measures library that have the same intervention category, the same intervention target, or meet the alternative matching parameters, and writes the alternative candidate measures into the candidate intervention set to form an alternative candidate intervention record.

[0057] The rule indexing module re-invokes the set of rules for conflicting alternative goals based on the intervention objects in the alternative candidate intervention records and the constraint categories in the constraint set.

[0058] The conflict detection module re-matches the intervention parameters in the alternative candidate intervention record, the structured field values ​​in the constraint set, and the triggering conditions in the alternative target conflict rule set to generate the alternative target conflict rule hit result.

[0059] When the result of the alternative target conflict rule is empty, the disposal execution module performs a retention mark writing on the alternative candidate intervention record; when the result of the alternative target conflict rule is not empty, it reads the corresponding disposal action field, disposal parameter field and disposal priority field, and performs execution condition writing, execution parameter adjustment, candidate measure deletion mark writing or review mark writing on the alternative candidate intervention record.

[0060] Furthermore, the closed-loop re-verification module generates frequency adjustment requests, duration adjustment requests, or execution condition adjustment requests based on compliance data, generates execution parameter adjustment requests based on physiological monitoring feedback data, and generates candidate measure deactivation requests, execution parameter downgrade requests, alternative candidate call requests, or review mark requests based on abnormal event records.

[0061] The closed-loop re-verification module converts the above request into adjusted candidate intervention measures and records the adjusted candidate intervention measures as adjusted candidate intervention records.

[0062] The rule indexing module re-invokes the target conflict rule set based on the intervention object in the adjusted candidate intervention record and the constraint category in the constraint set.

[0063] The conflict detection module matches the intervention parameters in the adjusted candidate intervention record, the structured field values ​​in the constraint set, and the trigger conditions in the re-called target conflict rule set to generate the adjusted target conflict rule hit result.

[0064] When the result of the adjusted target conflict rule is empty, the disposal execution module performs a retention mark writing on the adjusted candidate intervention record; when the result of the adjusted target conflict rule is not empty, it reads the disposal action field, disposal parameter field and disposal priority field in the adjusted target conflict rule, and performs execution condition writing, execution parameter adjustment, candidate measure deletion mark writing, alternative measure invocation or review mark writing on the adjusted candidate intervention record.

[0065] The intervention package generation module writes the reprocessed candidate intervention records, which have not been marked for deletion, into the corresponding intervention items in the lifestyle intervention package.

[0066] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention uses a candidate intervention set generation module to uniformly write multi-source candidate lifestyle interventions such as diet, exercise, sleep, psychology, social interaction, and medication adherence into the candidate intervention record. It also uses a constraint set generation module to convert disease diagnosis information, medication information, and comprehensive geriatric assessment information into constraint sets, so that candidate interventions from different sources enter a unified field-based processing flow, reducing inconsistencies in the handling of candidate measures caused by separate processing by different intervention modules.

[0067] 2. This invention establishes rule index keys based on the intervention objects in the candidate intervention records and the constraint categories in the constraint sets through the rule index module, and reads the target conflict rule set from the conflict rule base. This reduces the process of reading and matching rules that are irrelevant to the current candidate intervention records by the conflict detection module, and improves the targeting of rule calls when batch processing candidate intervention records.

[0068] 3. This invention generates target conflict rule hit results through a conflict detection module, and the disposal execution module performs actions such as writing retention flags, writing execution conditions, adjusting execution parameters, writing deletion flags, calling alternative measures, or writing review flags on candidate intervention records based on the disposal action field, disposal parameter field, and disposal priority field. This reduces the situation where the disposal parameters written to the same candidate intervention record are inconsistent when multiple rules are hit at the same time.

[0069] 4. This invention, through re-verification of alternative candidate intervention records and re-verification before closed-loop update, ensures that both alternative candidate measures and adjusted candidate intervention records formed by feedback adjustments re-enter the rule index, trigger condition matching, and disposal writing process, thereby preventing alternative measures or closed-loop adjustment results from bypassing the original constraint matching and disposal control process. Attached Figure Description

[0070] Figure 1 This is a schematic diagram of the system module structure of the present invention; Figure 2 This is a schematic diagram of the candidate intervention recording and constraint set generation process of the present invention; Figure 3 This is a schematic diagram of the conflict rule index and triggering condition matching process of the present invention; Figure 4 This is a schematic diagram of the process for generating the disposal and intervention package of the present invention; Figure 5 This is a schematic diagram of the alternative measures call and re-verification process of the present invention; Figure 6 This is a schematic diagram of the closed-loop re-verification process of the present invention; Figure 7 This is a schematic diagram of the log recording structure of the present invention. Detailed Implementation

[0071] To make the above-mentioned objectives, features, and advantages of the present invention more apparent and understandable, the present invention will be further described below in conjunction with the accompanying drawings and embodiments. It should be understood that the following embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the scope of protection of the present invention. The field names, field values, number of rules, rule parameters, threshold forms, terminal types, database types, and index types involved in each embodiment can be equivalently replaced according to the actual deployment scenario without changing the process of candidate intervention recording, rule indexing, trigger condition matching, disposal field reading, and candidate intervention record writing.

[0072] It should be noted that this invention is used for information processing of structured disease diagnosis information, medication information, comprehensive geriatric assessment information, and candidate lifestyle interventions.

[0073] The disease diagnosis information, medication information, comprehensive geriatric assessment information, physiological monitoring feedback data, and abnormal event records received by this invention are all used as input fields that have undergone structured processing to participate in the rule indexing, trigger condition matching, and writing of the treatment fields of candidate intervention records; the system does not change the original disease diagnosis information and medication information, nor does it generate output results that replace medical diagnosis and treatment conclusions.

[0074] When receiving structured input fields, the data receiving module reads the source identifier, authorization identifier, and timestamp simultaneously and writes the above information into the input field log; input fields that do not carry an authorization identifier or source identifier will not enter the subsequent rule indexing and trigger condition matching process.

[0075] Multi-source candidate lifestyle interventions are converted into candidate intervention records. Disease diagnosis information, medication information, and comprehensive geriatric assessment information are converted into constraint sets. A rule index key is formed based on the intervention target in the candidate intervention record and the constraint category in the constraint set. The target conflict rule set is read from the conflict rule base. The intervention parameters in the candidate intervention record and the structured field values ​​in the constraint set are matched with the trigger condition fields. Based on the target conflict rule hit result, the treatment action field, treatment parameter field, and treatment priority field are read. Execution conditions are written, execution parameters are adjusted, deletion markers are written, and alternative measure invocation or review markers are written to the candidate intervention record. Replacement candidate intervention records and adjusted candidate intervention records formed by closed-loop updates must re-enter the rule indexing, trigger condition matching, and treatment writing process before being written into the lifestyle intervention package by the intervention package generation module.

[0076] Example 1, referring to Figures 1-4 This is the first embodiment of the present invention, which provides a personalized lifestyle intelligent intervention system for elderly people with comorbidities.

[0077] Specifically, this embodiment is applied to a scenario where a community-based elderly care service center generates an initial lifestyle intervention package for elderly users with comorbidities.

[0078] The community elderly care service center has been connected to the health record system, the comprehensive elderly assessment system, the user terminal, the caregiver terminal, and the management terminal.

[0079] The system is deployed on a community health management server, which includes a processor, memory, and a database. The memory stores the program instructions executed by the processor, and the database stores field mapping tables, conflict rule bases, alternative measure bases, closed-loop adjustment rule tables, candidate intervention records, lifestyle intervention packages, and processing logs. The rule index keys are implemented using database composite indexes, hash indexes, or key-value indexes.

[0080] When the processor executes program instructions, it calls the data receiving module, constraint set generation module, candidate intervention set generation module, rule indexing module, conflict detection module, treatment execution module, intervention package generation module, and closed-loop re-verification module to record, index, match, write, and generate intervention packages for candidate lifestyle intervention measures.

[0081] This system is an information system design, information system integration, and information processing and storage support system for lifestyle intervention scenarios involving comorbidities in the elderly. It utilizes servers, databases, field mapping tables, conflict rule bases, alternative measure bases, closed-loop adjustment rule tables, candidate intervention records, lifestyle intervention packages, and processing logs to uniformly receive, store, index, match, and write structured fields from multiple sources. Data connections between the health record system, the comprehensive elderly assessment system, the user terminal, the caregiver terminal, and the management terminal form a system integration structure. The user terminal, caregiver terminal, and management terminal are respectively used for displaying lifestyle intervention packages, providing execution condition prompts, receiving review markers, maintaining conflict rules, and managing rule versions, forming a terminal interaction and status feedback process for candidate intervention processing results.

[0082] Reference Figure 1 The system includes a data receiving module, a constraint set generation module, a candidate intervention set generation module, a rule indexing module, a conflict detection module, a disposal execution module, an intervention package generation module, and a closed-loop re-verification module.

[0083] Specifically, the data receiving module receives structured disease diagnosis information, medication information, comprehensive geriatric assessment information, and candidate lifestyle interventions. Disease diagnosis information includes fields for disease name, disease code, disease status, and entry time; medication information includes drug category label, medication time, and medication reminder; comprehensive geriatric assessment information includes fields for frailty status, fall risk, cognitive status, nutritional risk, daily living abilities, and care resources; and candidate lifestyle interventions include dietary interventions, exercise interventions, sleep interventions, psychological interventions, social interventions, and medication adherence interventions.

[0084] In another application scenario, comprehensive assessment information for the elderly also includes cognitive function status fields, neurological disease follow-up marker fields, or brain disease-related event record fields provided by external medical information systems, follow-up systems, or standardized assessment tools.

[0085] The above fields serve as structured input fields for the constraint set generation module, used to convert them into execution complexity constraints, task execution constraints, or review trigger constraints. After being written into the constraint set, these fields only participate in rule indexing and trigger condition matching as structured fields for execution complexity constraints, task execution constraints, or review trigger constraints. The disposal execution module writes execution conditions, review markers, or adjustment parameters to candidate intervention records based on the matching results.

[0086] Specifically, in this embodiment, the target user is the recipient of community-based elderly care services. The structured input data received by the data receiving module includes: disease diagnosis information including markers for abnormal glucose metabolism, hypertension, and limited knee joint function; medication information including drug category markers, medication time fields, and medication reminder fields; and comprehensive elderly assessment information including a high fall risk field, a mild decline in cognitive status, a moderate nutritional risk field, partial assistance in daily living skills, and care resources with caregivers available during the day.

[0087] Candidate lifestyle interventions include a 30-minute walk after meals, low-sodium diet reminders, standing resistance training, fixed bedtime reminders, weekly community activity reminders, and medication check-in reminders.

[0088] Reference Figure 2 The system executes the candidate intervention recording and constraint set generation process, which specifically includes the following steps: S101, the data receiving module performs structured field reception: the data receiving module receives disease diagnosis information and medication information from the health record system, receives comprehensive elderly assessment information from the comprehensive elderly assessment system, receives candidate lifestyle intervention measures from the candidate intervention generation end, and writes the reception time, source identifier and authorization identifier into the input field log.

[0089] Among them, the candidate intervention generation end is an external candidate measure generation module, management end configuration module, or third-party lifestyle intervention measure library connected to the system.

[0090] S102, the data receiving module performs field verification: field verification includes field name verification, field type verification, field enumeration value verification, field null value verification, source identifier verification, and timestamp verification.

[0091] The data receiving module determines the field verification result based on whether the field exists, whether the field value belongs to the preset enumeration range, whether the field type conforms to the preset type, and whether the field has an authorized source identifier and an entry timestamp.

[0092] For example, the fall risk field can be configured with values ​​ranging from low to high; the cognitive status field can be configured with values ​​ranging from normal to slightly impaired to moderately impaired; and the care resources field can be configured with values ​​ranging from none to caregivers available during the day to caregivers available all day. When the fall risk field is empty, the data receiving module does not infer fall risk but instead writes a fall risk field missing flag.

[0093] S103, The constraint set generation module performs constraint set transformation: The constraint set generation module reads the field mapping table, converts disease diagnosis information into disease constraint subsets, medication information into medication constraint subsets, and geriatric comprehensive assessment information into geriatric comprehensive assessment constraint subsets.

[0094] The disease constraint subset stores disease constraint records converted from disease diagnosis information; the medication constraint subset stores medication constraint records converted from medication information; and the geriatric comprehensive assessment constraint subset stores functional status constraint records converted from geriatric comprehensive assessment information.

[0095] The above conversion is completed based on the field mapping table, and is not subject to temporary medical judgments by the system during operation.

[0096] S104, The constraint set generation module performs the mapping of the geriatric comprehensive assessment fields: The constraint set generation module converts the frailty status field into frailty constraints, the fall risk field into fall constraints, the cognitive status field into execution complexity constraints, the nutritional risk field into dietary execution constraints, the daily living ability field into task execution constraints, and the care resource field into care assistance constraints, and writes the conversion results into the geriatric comprehensive assessment constraint subset.

[0097] For example, when the fall risk field is high, the system writes a fall constraint; when the cognitive state field is slightly decreased, the system writes an execution complexity constraint; when the daily living ability field is partially assisted, the system writes a task execution constraint; and when the care resource field is that there is a caregiver during the day, the system writes a care assistance constraint.

[0098] S105, The candidate intervention set generation module performs candidate intervention recording: The candidate intervention set generation module receives candidate lifestyle interventions from multiple sources and writes each candidate lifestyle intervention into a candidate intervention record, and the candidate intervention set is formed by multiple candidate intervention records.

[0099] The candidate intervention record includes candidate source, intervention category, intervention target, execution frequency, execution duration, execution intensity, execution time, target parameters, intervention parameters, and integrity markers.

[0100] For example, after a 30-minute walk after a meal is written into the candidate intervention record, the candidate source is the exercise intervention module, the intervention category is exercise intervention, the intervention target is walking exercise, the execution frequency is once a day, the execution duration is 30 minutes, the execution intensity is moderate, and the execution time is after dinner; after a fixed bedtime reminder is written into the candidate intervention record, the candidate source is the sleep intervention module, the intervention category is sleep intervention, the intervention target is sleep time reminder, the execution time is 21:30, and the intervention parameters include the number of reminder steps and the number of reminders.

[0101] S106, the candidate intervention set generation module performs integrity marking: When a candidate lifestyle intervention is missing information such as the intervention target, duration, intensity, or target parameters, the candidate intervention set generation module does not add subjective content but instead writes a corresponding field missing marker into the candidate intervention record. The field missing marker serves as input for the subsequent conflict detection module to match field missing conditions.

[0102] The execution intensity limit, frequency limit, duration limit, restriction intensity limit, step number threshold, and feedback threshold involved in the above embodiments are not generated temporarily by the system during operation.

[0103] The management system pre-writes thresholds, enumeration boundaries, comparison operators, and applicable fields into the conflict rule base based on publicly released clinical guidelines, drug instructions, pharmacology databases, standardized comprehensive geriatric assessment tools, or standardized assessment forms whose source name, version number, applicable fields, and activation status are recorded by the management system.

[0104] Standardized comprehensive assessment tools for the elderly include the Fried frailty assessment, the Morse fall risk assessment, the nutritional risk screening scale, and the activities of daily living assessment scale.

[0105] The aforementioned thresholds and enumeration boundaries are not temporarily entered or modified by the operator during a single user processing session.

[0106] The conflict rule base stores the source name, source version, entry time, applicable fields, enabled status, and rule version identifier for each threshold record.

[0107] When performing matching, the conflict detection module reads the field values ​​of the thresholds, comparison operators, enumeration boundaries, and candidate intervention records or constraint sets that have been written into the conflict rule base, and outputs the field matching results or threshold comparison results; the handling execution module writes the corresponding handling fields according to the matching results.

[0108] S107, the system performs input missing handling. When any type of input data, such as disease diagnosis information, medication information, or comprehensive geriatric assessment information, is an empty set, the data receiving module writes a missing input category marker, and the constraint set generation module writes an incomplete constraint set construction marker.

[0109] The conflict detection module reads the constraint category field of each target conflict rule in the target conflict rule set. When a target conflict rule depends on a missing input category, the missing field condition is determined as a successful match, and a target conflict rule hit result is generated.

[0110] The action execution module writes a review flag to candidate intervention records that depend on the missing input category based on the action field corresponding to the missing field condition. When there are output control parameters configured in the conflict rule base, the action execution module also writes an output blocking flag, and the intervention package generation module skips the corresponding candidate intervention record based on the flag.

[0111] After the management side completes the missing input data or writes the authorized default processing policy, the rule index module calls the target conflict rule set again, and the conflict detection module re-matches the trigger conditions.

[0112] Reference Figure 3 The system executes a conflict rule index and trigger condition matching process, which specifically includes the following steps: Before the rule index module reads the conflict rule base, the management side executes the conflict rule structured configuration process.

[0113] Specifically, the management terminal receives rule source information, applicable objects, applicable fields, comparison operators, threshold fields, enumeration boundaries, trigger condition fields, action fields, action parameter fields, and action priority fields, and writes the above content into the rule editing interface according to the preset field template to form a record of conflict rules to be added to the database.

[0114] The record of conflict rules to be added to the database includes rule identifier, rule version identifier, rule source identifier, constraint category field, constraint object field, intervention object field, trigger condition field, action field, action parameter field, and action priority field.

[0115] The management side performs field integrity checks and field type checks on the conflict rule records to be entered into the database. Field integrity checks include checking whether the rule identifier, rule version identifier, rule source identifier, constraint category field, constraint object field, intervention object field, trigger condition field, action field, and action parameter field are empty.

[0116] Field type validation includes validating whether the threshold field is a numeric field, whether the enumeration boundary belongs to the preset enumeration range, whether the comparison operator belongs to the preset comparison operator set, whether the trigger condition field can be parsed as a field matching condition, field missing condition, tag hit condition, threshold comparison condition or intervention parameter out-of-bounds condition, and whether the action field belongs to one of the following: condition writing action, parameter adjustment action, candidate measure deletion tag writing action, alternative measure invocation action or review tag writing action.

[0117] When a conflict rule record to be added passes field integrity and field type validation, the management system writes the record to the conflict rule database, along with its activation status, entry time, and rule version identifier. When a conflict rule record fails either field integrity or field type validation, the management system adds a rule configuration error flag to the record and prevents it from entering the callable rule set of the rule index module.

[0118] For example, the management system generates a conflict rule record based on the fall risk field in the fall risk assessment form and the execution intensity field in the exercise candidate intervention template. This conflict rule record has the following characteristics: constraint category field is fall constraint, constraint object field is fall risk field, intervention object field is walking exercise, trigger condition field is the fall risk field hitting a preset high-risk marker and the execution intensity field hitting a medium or high intensity enumeration value, action field is parameter adjustment action, and action parameter field is writing the execution intensity field to a low intensity enumeration value and splitting the continuous execution duration field into segmented execution duration fields. After performing field integrity and field type validation on this conflict rule record, the management system writes the conflict rule record to the conflict rule database.

[0119] S108, the rule index module reads the conflict rule library. The rule index module reads conflict rules from the conflict rule library that are enabled and have not been marked with a rule configuration error, as a set of callable rules. The conflict rule library stores multiple conflict rules. Each conflict rule includes a rule identifier, rule version identifier, rule source identifier, constraint category field, constraint object field, intervention object field, trigger condition field, action field, action parameter field, and action priority field.

[0120] Rule identifiers are used to distinguish different conflicting rules in the conflict rule base; rule version identifiers are used to record the version of the conflicting rule; rule source identifiers are used to record the source of the conflicting rule.

[0121] The constraint category field stores constraint category values ​​that match the constraint categories in the constraint set; the constraint object field stores constraint object values ​​that match the structured field values ​​in the constraint set; the intervention object field stores intervention object values ​​that match the intervention objects in the candidate intervention records; the trigger condition field stores field matching conditions, field missing conditions, flag hit conditions, threshold comparison conditions, or intervention parameter out-of-bounds conditions; the treatment action field stores condition writing actions, parameter adjustment actions, candidate measure deletion flag writing actions, alternative measure invocation actions, or review flag writing actions; the treatment parameter field stores writing parameters, adjustment parameters, alternative matching parameters, or review field parameters corresponding to the treatment action field; and the treatment priority field stores execution sorting parameters when multiple target conflict rules are hit simultaneously.

[0122] S109, The rule index module establishes the rule index key: The rule index module establishes the rule index key according to the constraint category field and the intervention object field, reads the intervention object from the candidate intervention record, reads the constraint category with written constraint value from the constraint set, and then forms the rule index key with the intervention object and the constraint category with written constraint value.

[0123] S110, the rule index module calls the target conflict rule set: the rule index module reads the target conflict rule set from the conflict rule base according to the rule index key.

[0124] For example, the candidate intervention record is a 30-minute walk after a meal, and the intervention target is walking exercise; the constraint set already includes fall restriction and care assistance restriction. The rule indexing module uses walking exercise plus fall restriction and walking exercise plus care assistance restriction as rule index keys to read the corresponding target conflict rule set from the conflict rule base.

[0125] S111, The conflict detection module reads the matching data: The conflict detection module reads the constraint object field and trigger condition field of each target conflict rule in the target conflict rule set, and reads the structured field values ​​in the constraint set and the intervention parameters in the candidate intervention records.

[0126] S112, The conflict detection module performs trigger condition matching: The conflict detection module performs equality matching between structured field values ​​and field matching conditions, matches empty fields with field missing conditions, performs tag matching between structured field values ​​and tag hit conditions, compares structured field values ​​with threshold comparison conditions, and compares intervention parameters with upper and lower limits of intervention parameter out-of-bounds conditions.

[0127] S113, The conflict detection module generates target conflict rule hit results: When the structured field value corresponding to the constraint object field successfully matches the trigger condition field, or when the intervention parameter in the candidate intervention record successfully matches the trigger condition field, the conflict detection module generates target conflict rule hit results. The hit results include the candidate intervention record identifier, the target conflict rule identifier, the hit field, the hit condition, the hit time, and the rule source identifier.

[0128] Specifically, for the candidate intervention record of walking 30 minutes after a meal, the conflict detection module reads a high value for the fall risk field, a medium intensity value for the execution intensity in the candidate intervention record, and the trigger condition in the first rule record as a high fall risk field and a medium or high intensity execution intensity. Since the field values ​​meet the trigger conditions, the conflict detection module generates a target conflict rule hit result. If the care resource field is empty, the second rule record also generates a target conflict rule hit result; if the care resource field indicates a caregiver during the day, the second rule record does not generate a target conflict rule hit result.

[0129] Specifically, the triggering conditions, upper threshold, lower threshold, handling actions, and handling parameters in the conflict rule base are pre-saved by the management end in the form of structured rule files or database records.

[0130] The values ​​for execution intensity, constraint intensity, care resource status, cognitive status, and daily living ability status are structured enumeration values ​​from candidate intervention records or constraint sets.

[0131] The conflict detection module and the handling execution module process the structured fields in the candidate intervention records, constraint sets, and target conflict rules. Processing conditions include field existence status, field null value status, matching results between field values ​​and preset markers, comparison results between field values ​​and preset upper or lower limits, and consistency of values ​​when multiple handling actions are written to the same parameter location. All of these processing conditions are based on the structured fields and rule fields already written into the system.

[0132] The trigger condition field can store single-field conditions or combined conditions consisting of multiple structured fields.

[0133] The treatment priority field is only used as a sorting parameter for computer execution when multiple target conflict rules are hit simultaneously, and does not indicate the level of medical risk or the severity of the disease; the treatment priority field, along with the treatment action field and treatment parameter field, is pre-stored in the conflict rule base and recorded with the rule version identifier.

[0134] In this embodiment, the conflict rule base stores the following rule records: The first rule record stores the fall constraint in the constraint category field, the fall risk field in the constraint object field, the walking movement in the intervention object field, the fall risk field in the trigger condition field, the fall risk field in the execution intensity field, the parameter adjustment action in the disposal action field, the disposal parameter field in the disposal parameter field, the disposal priority field in the disposal priority field, and the second priority in the disposal priority field.

[0135] The second rule record stores the care assistance constraint in the constraint category field, the care resource field in the constraint object field, the walking movement in the intervention object field, the fall risk field in the trigger condition field when the fall risk field is high and the care resource field is empty, the treatment action field in the treatment action field in the review mark writing action, the treatment parameter field in the treatment parameter field in the review mark and the reason for missing care conditions, and the treatment priority field in the treatment priority field in the first priority field.

[0136] The third rule record's constraint category field stores the execution complexity constraint, the constraint object field stores the cognitive state field, the intervention object field stores the sleep time reminder, the trigger condition field stores the cognitive state field as slightly declining and the number of reminder steps in the candidate intervention record is greater than the preset number of steps threshold, the treatment action field stores the condition writing action, the treatment parameter field stores the caregiver's synchronous reminder conditions and simplified execution step parameters, and the treatment priority field stores the third priority.

[0137] The fourth rule record's constraint category field stores the dietary execution constraint, constraint object field stores the nutritional risk field, intervention object field stores the dietary restriction measures, trigger condition field stores the nutritional risk field as medium or high and the dietary restriction intensity exceeds the preset upper limit, treatment action field stores the review mark writing action, treatment parameter field stores the nutritional review mark, and treatment priority field stores the first priority.

[0138] The fifth rule record's constraint category field stores data integrity constraints, constraint object field stores key assessment fields, intervention object field stores dietary intake adjustments, trigger condition field stores key assessment fields associated with the intervention object that are empty, treatment action field stores the action of writing the review mark, treatment parameter field stores the review mark for missing fields, and treatment priority field stores the first priority.

[0139] The sixth rule record's constraint category field stores task execution constraints, constraint object field stores daily living ability fields, intervention object field stores standing resistance training, trigger condition field stores daily living ability fields as partial assistance and execution intensity exceeding the preset intensity limit, treatment action field stores candidate measure deletion mark writing action and alternative measure invocation action, treatment parameter field stores deletion mark field value and alternative measure matching parameter, alternative measure matching parameter is seated resistance training or elastic band low-intensity training, and treatment priority field stores the second priority.

[0140] Reference Figure 4 The system executes the disposal and intervention package generation process, which specifically includes the following steps: S114, the handling execution module determines whether the hit result is empty: the handling execution module reads the target conflict rule hit result. When the target conflict rule hit result is empty, the module performs a retention mark writing on the candidate intervention record. In the case where the retention mark writing corresponds to the target conflict rule hit result being empty, the handling execution module does not read the non-existent handling action field.

[0141] S115, The handling execution module reads the handling fields: When the target conflict rule hit result is not empty, the handling execution module reads the handling action field, handling parameter field and handling priority field in the target conflict rule corresponding to the target conflict rule hit result, and sorts the execution according to the handling priority field.

[0142] S116, Execution Module Writes Records: The Execution Module writes execution conditions, adjusts execution parameters, writes candidate measure deletion markers, and writes alternative measure invocation or review markers to candidate intervention records according to the action field and parameter field.

[0143] Specifically, when the first rule record is hit after walking for 30 minutes after a meal, the handling execution module reads the parameters, adjusts the action and corresponding handling parameters, adjusts the execution intensity from medium to low intensity, and adjusts the execution duration from continuous 30 minutes to a segmented execution format of 10 minutes each time, 3 times a day.

[0144] When a 30-minute walk after a meal triggers the second rule record, and the second rule record has a higher priority than the first rule record, the action execution module will prioritize writing the review flag and record the reason for the missing care conditions.

[0145] S117, The handling execution module performs multi-rule merging: When the handling execution module hits multiple target conflict rules in the same candidate intervention record, it sorts the multiple target conflict rules according to the handling priority field, and merges the handling parameters that can be written to the same candidate intervention record at the same time into the candidate intervention record.

[0146] For example, if one rule is written for low-intensity execution and another rule is written for daytime execution, and the two treatment parameters are written in different parameter positions, the treatment execution module will merge the two treatment parameters and write them into the candidate intervention record.

[0147] When the action field of a high-priority target conflict rule contains a candidate measure deletion mark writing action, the action execution module writes the deletion mark to the candidate intervention record; subsequent execution condition writing actions or execution parameter adjustment actions for the same candidate intervention record are only written to the action execution log and are no longer written to the executable parameter field.

[0148] S118, Execution Module Conflict Marker: When the execution module writes the same parameter position in multiple action fields and the values ​​are inconsistent, or when the execution objects of multiple action fields are mutually exclusive, a review mark is written to the candidate intervention record.

[0149] For example, one rule requires adjusting a certain sport to low intensity, and another rule requires writing a deletion mark for the same sport. The system sorts the rules according to the priority field. If, after sorting, there are still inconsistent values ​​for the same parameter or mutually exclusive execution objects, the execution module writes a review mark and records the conflict rule identifier.

[0150] S119, Disposal Execution Module Execution Deletion Mark Control: When the disposal action field contains the candidate measure deletion mark writing action, the disposal execution module writes a deletion mark that will not be included in the lifestyle intervention package to the candidate intervention record.

[0151] The deletion flag, as a record-level control field, is used to control the intervention package generation module to skip the corresponding candidate intervention record.

[0152] Candidate intervention records marked with deletion are retained in the system for log tracing, review, and alternative matching.

[0153] S120, The intervention package generation module generates a lifestyle intervention package: The intervention package generation module reads the retention flag, execution condition field, execution parameter field, deletion flag field, alternative measure flag field and review flag field from the candidate intervention record after it has been processed by the treatment execution module.

[0154] The intervention package generation module writes candidate intervention records that have not been marked for deletion into the lifestyle intervention package.

[0155] For candidate intervention records that have been marked for deletion, the intervention package generation module does not write them into a lifestyle intervention package.

[0156] For candidate intervention records marked with a review flag, the intervention package generation module writes a review prompt field or sends the candidate intervention record to the review end.

[0157] The review end is the review interface in the management end, or the audit terminal connected to the management end, used to receive review marks, conflict rule identifiers, and candidate intervention records.

[0158] Specifically, the lifestyle intervention package includes intervention categories, intervention items, implementation frequency, implementation duration, implementation intensity, implementation time, implementation conditions, implementation suspension conditions, review trigger conditions, feedback thresholds, alternative measure markers, and review markers.

[0159] The execution pause conditions, review trigger conditions, and feedback thresholds are written by preset fields in the conflict rule base, closed-loop adjustment rule table, or disposal parameter fields.

[0160] After treatment, the lifestyle intervention package can include post-meal low-intensity segmented walking, 10 minutes each time, 3 times a day, to be performed during the daytime, in the presence of a caregiver; for candidate intervention records of standing resistance training marked for deletion, the intervention package generation module will not include the item in the lifestyle intervention package.

[0161] Example 2, refer to Figures 1-5 This is the second embodiment of the present invention. The difference between this embodiment and the first embodiment is that this embodiment further provides an application scenario for invoking alternative measures and re-verifying alternative candidate measures.

[0162] Specifically, this embodiment is applied to the scenario of alternative exercise interventions in home rehabilitation services. In this scenario, the candidate intervention set contains exercise candidate intervention records corresponding to standing resistance training, and the constraint set has already written fall constraints and task execution constraints. The system performs deletion marking, generation of alternative candidate records, and re-verification processing on the candidate intervention record according to the goal conflict rules.

[0163] Reference Figure 5 The target users' comprehensive elderly assessment information includes a high fall risk field, a partially assisted daily living ability field, and a care resource field indicating that they have caregivers during the day.

[0164] The candidate intervention set includes candidate intervention records corresponding to standing resistance training. The candidates are sourced from the exercise intervention module, the intervention category is exercise intervention, the intervention target is standing resistance training, the execution frequency is 3 times a week, the execution duration is 20 minutes, and the execution intensity is moderate.

[0165] Reference Figure 5 This embodiment executes the alternative measure invocation and re-verification process, specifically including the following steps: S201, the rule indexing module calls the target conflict rule set corresponding to the original candidate intervention record. Based on the intervention object in the candidate intervention record corresponding to the standing resistance training and the task execution constraints in the constraint set, the rule indexing module calls the target conflict rule set corresponding to the standing resistance training.

[0166] S202, the conflict detection module performs trigger condition matching on the original candidate intervention record.

[0167] The conflict detection module reads the daily living skills field, the fall risk field, and the execution intensity field from the candidate intervention records, and matches the field values ​​with the trigger condition field of any target conflict rule.

[0168] S203, the disposal execution module writes a deletion mark to the original candidate intervention record.

[0169] When the trigger condition field of any target conflict rule is successfully matched, the disposal execution module reads the disposal action field and the disposal parameter field, writes a deletion mark to the candidate intervention record corresponding to the standing resistance training, and reads the alternative measure matching parameter in the disposal parameter field.

[0170] S204, The handling execution module reads the alternative measures library.

[0171] The alternative measures library stores alternative measure identifiers, intervention categories, intervention targets, applicable conditions, prohibited conditions, source identifiers, and version identifiers.

[0172] The source identifier is used to record that the alternative candidate measures are derived from management configuration, standardized lifestyle intervention templates of institutions, publicly available guide entries, or authorized imported intervention measure libraries; the version identifier is used to record the updated version of the alternative candidate measures.

[0173] S205, the handling execution module calls the alternative candidate measure.

[0174] The treatment execution module reads the intervention category and intervention object from the candidate intervention record marked with deletion, and reads the alternative matching parameter from the treatment parameter field; the treatment execution module calls alternative candidate measures with the same intervention category, the same intervention object, or that meet the alternative matching parameter from the alternative measures library.

[0175] For example, the alternative measures library stores seated low-intensity resistance training with elastic bands, seated leg raises, and chair-assisted low-intensity resistance training. The treatment execution module calls the seated low-intensity resistance training with elastic bands based on the type of exercise intervention, the resistance training object, and the alternative matching parameters.

[0176] S206, The disposal execution module generates alternative candidate intervention records.

[0177] The treatment execution module writes the low-intensity sit-and-hold elastic band training into the candidate intervention set and forms an alternative candidate intervention record.

[0178] This alternative candidate intervention record includes the intervention category, target population, frequency of implementation, duration of implementation, intensity of implementation, implementation conditions, and intervention parameters. The alternative candidate intervention record is not directly incorporated into the lifestyle intervention package.

[0179] S207, the rule index module re-invokes the set of rules for the alternative target conflict.

[0180] The rule indexing module re-invokes the set of rules for conflicting alternative goals based on the intervention objects in the alternative candidate intervention records and the constraint categories in the constraint set.

[0181] S208, The conflict detection module performs trigger condition matching for alternative candidate intervention records.

[0182] The conflict detection module re-matches the intervention parameters in the alternative candidate intervention record, the structured field values ​​in the constraint set, and the triggering conditions in the alternative target conflict rule set to generate the alternative target conflict rule hit result.

[0183] S209, The disposal execution module processes alternative candidate intervention records.

[0184] When the result of the alternative target conflict rule is empty, the disposal execution module performs a retention mark writing on the alternative candidate intervention record; when the result of the alternative target conflict rule is not empty, it reads the corresponding disposal action field, disposal parameter field and disposal priority field, and performs execution condition writing, execution parameter adjustment, candidate measure deletion mark writing or review mark writing on the alternative candidate intervention record.

[0185] S210, the intervention package generation module writes the alternative candidate intervention record.

[0186] The intervention package generation module writes alternative candidate intervention records that have been reprocessed and have not been marked for deletion into the lifestyle intervention package.

[0187] For example, if a low-intensity sit-and-hold resistance band training session is not marked for deletion after re-verification, the intervention package generation module will write the alternative candidate intervention record into the lifestyle intervention package and write execution conditions such as sit-and-hold, 10 minutes per session, 3 times per week, and execution in the presence of a caregiver.

[0188] After alternative candidate measures are written into the candidate intervention set and form alternative candidate intervention records, they still need to go through the rule indexing, trigger condition matching, and disposal writing process again. If a deletion mark is not written, the intervention package generation module will then write them into the lifestyle intervention package. The rest of the structure is the same as that in Example 1.

[0189] Example 3, referring to Figures 1-7 This is the third embodiment of the present invention. The difference between this embodiment and the first embodiment is that this embodiment further provides an application scenario for re-verification before the closed-loop update of the lifestyle intervention package.

[0190] Specifically, this embodiment is applied to the feedback and adjustment scenario after home-based elderly care users implement lifestyle intervention packages.

[0191] After users, caregivers, or monitoring devices upload compliance data, physiological monitoring feedback data, or abnormal event records, the system does not directly overwrite the original intervention items in the lifestyle intervention package. Instead, it first generates an intervention package adjustment request, converts the intervention package adjustment request into an adjusted candidate intervention record, and then re-enters the rule index, trigger condition matching, and disposal writing process.

[0192] In this embodiment, the lifestyle intervention package already includes a low-intensity segmented walk after meals, 10 minutes each time, 3 times a day, performed during the daytime, and performed in the presence of a caregiver.

[0193] The closed-loop revalidation module receives compliance data, physiological monitoring feedback data, and abnormal event records corresponding to the lifestyle intervention package.

[0194] Compliance data includes the number of times a task is completed, the number of times a task is skipped, the number of times a task is delayed, and user confirmation records; physiological monitoring feedback data includes structured feedback fields such as self-measured heart rate, self-measured blood pressure, self-measured blood glucose, steps, sleep duration, and fatigue score; abnormal event records include records of falls, hypoglycemia, abnormal blood pressure, significant reduction in sleep, or discomfort after exercise.

[0195] The closed-loop re-verification module reads the field values ​​from the above data and generates frequency adjustment requests, duration adjustment requests, execution condition adjustment requests, execution parameter adjustment requests, candidate measure discontinuation requests, execution parameter downgrade requests, alternative candidate call requests, or review mark requests based on the closed-loop adjustment rule table. The closed-loop re-verification module does not change the disease diagnosis information or medication information; it only generates adjustment requests and performs re-verification processing on the candidate intervention records in the lifestyle intervention package.

[0196] Specifically, the field values ​​in the physiological monitoring feedback data are used to match the feedback thresholds, comparison operators, and treatment parameters in the closed-loop adjustment rule table or conflict rule base, and generate execution parameter adjustment requests, execution condition adjustment requests, or review mark requests accordingly; the closed-loop re-verification module does not change the disease diagnosis information or medication information.

[0197] Reference Figure 6 This embodiment performs a closed-loop re-verification process, which specifically includes the following steps: S301, the closed-loop revalidation module receives feedback data. The closed-loop revalidation module receives compliance data, physiological monitoring feedback data, and abnormal event records corresponding to the lifestyle intervention package, and writes the feedback data source, timestamp, and associated intervention project identifier into the closed-loop revalidation log.

[0198] S302, the closed-loop re-verification module generates an adjustment request based on compliance data.

[0199] The closed-loop re-verification module generates frequency adjustment requests, duration adjustment requests, or execution condition adjustment requests based on compliance data.

[0200] For example, when a target user completes only a number of walking tasks below a preset completion threshold within N consecutive statistical periods, the closed-loop revalidation module generates a frequency adjustment request; where N and the preset completion threshold are pre-written into the closed-loop adjustment rule table. The closed-loop revalidation module only reads the task completion count, statistical period, candidate intervention record identifier, and threshold field from the compliance data and generates a corresponding adjustment request based on the field comparison results.

[0201] S303, the closed-loop recalibration module generates an adjustment request based on physiological monitoring feedback data.

[0202] When the post-exercise heart rate field in the physiological monitoring feedback data meets the threshold comparison condition in the closed-loop adjustment rule table, the closed-loop re-verification module generates an execution parameter adjustment request and converts the request into an adjusted candidate intervention record.

[0203] S304, The closed-loop re-verification module generates an adjustment request based on the abnormal event record.

[0204] The closed-loop re-verification module generates candidate measure deactivation requests, execution parameter downgrade requests, alternative candidate call requests, or review mark requests based on the abnormal event records.

[0205] For example, when the abnormal event log includes a fall event log, the closed-loop re-verification module generates a candidate measure deactivation request or a review flag request.

[0206] S305, the closed-loop re-verification module generates adjusted candidate intervention records.

[0207] The closed-loop re-verification module converts frequency adjustment requests, duration adjustment requests, execution condition adjustment requests, execution parameter adjustment requests, candidate measure deactivation requests, execution parameter downgrade requests, alternative candidate call requests, or review mark requests into adjusted candidate intervention measures, and records the adjusted candidate intervention measures as adjusted candidate intervention records.

[0208] For example, a frequency adjustment request could be converted into a post-meal low-intensity segmented walk, 10 minutes each time, once a day, and form an adjusted candidate intervention record.

[0209] S306, the rule index module re-invokes the target conflict rule set.

[0210] The rule indexing module re-invokes the target conflict rule set based on the intervention object in the adjusted candidate intervention record and the constraint category in the constraint set.

[0211] S307, the conflict detection module re-matches the trigger conditions.

[0212] The conflict detection module matches the intervention parameters in the adjusted candidate intervention record, the structured field values ​​in the constraint set, and the trigger conditions in the re-called target conflict rule set to generate the adjusted target conflict rule hit result.

[0213] S308, the handling and execution module processes and adjusts candidate intervention records.

[0214] When the result of the adjusted target conflict rule is empty, the disposal execution module writes a retention mark to the adjusted candidate intervention record; when the result of the adjusted target conflict rule is not empty, the disposal execution module reads the disposal action field, disposal parameter field and disposal priority field in the adjusted target conflict rule, and writes the execution conditions, adjusts the execution parameters, writes the candidate measure deletion mark, and writes the alternative measure call or review mark to the adjusted candidate intervention record.

[0215] S309, the intervention package generation module updates the corresponding intervention items.

[0216] The intervention package generation module writes the reprocessed, adjusted candidate intervention records (without deletion markers) into the corresponding intervention items in the lifestyle intervention package. The adjusted candidate intervention records do not directly overwrite the original intervention items; instead, they are written into the corresponding intervention items after reprocessing.

[0217] Reference Figure 7 The system generates and processes log records.

[0218] Specifically, the processing log structure includes input field logs, candidate intervention record logs, rule index logs, conflict detection logs, disposal execution logs, alternative measure re-verification logs, and closed-loop re-verification logs.

[0219] The input field log saves the input category, field name, field type, field value summary, source identifier, authorization identifier, and timestamp.

[0220] The candidate intervention log saves the candidate intervention record identifier, candidate source, intervention category, intervention target, execution frequency, execution duration, execution intensity, execution time, target parameters, and intervention parameters.

[0221] The rule index log stores the rule index key, the intervention object, the constraint category of the constraint value that has been written, and the identifier of the target conflict rule that has been read.

[0222] The conflict detection log saves the target conflict rule identifier, constraint object field, trigger condition field, structured field values ​​involved in the matching, intervention parameters involved in the matching, matching result, and hit time.

[0223] Specifically, the processing execution log saves the processing action field, processing parameter field, processing priority field, write parameter position, write result, deletion mark, review mark and execution time.

[0224] The alternative measure revalidation log saves the original candidate intervention record identifier, the alternative candidate measure identifier, the alternative matching parameter, the alternative target conflict rule identifier, and the alternative target conflict rule hit result.

[0225] The closed-loop revalidation log saves compliance data identifiers, physiological monitoring feedback data identifiers, abnormal event record identifiers, intervention package adjustment request types, adjusted candidate intervention record identifiers, rematch results, and lifestyle intervention package update results.

[0226] The processing log is used to record the processes of field reading, rule indexing, trigger condition matching, disposal field writing, replacement re-verification, and closed-loop re-verification, and is used to trace the processing chain of candidate intervention records in the future.

[0227] The adjusted candidate intervention record will not directly overwrite the original intervention item in the lifestyle intervention package. Instead, it will re-enter the rule index, trigger condition matching, and disposal writing process. After reprocessing and without writing a deletion mark, the intervention package generation module will then write the adjusted candidate intervention record into the corresponding intervention item in the lifestyle intervention package. The rest of the structure is the same as that in Example 1.

[0228] In a processing method that does not use rule index keys, the system iterates through all rules in the conflict rule base for each candidate intervention record; as the number of candidate intervention records and conflict rules increases, the system needs to repeatedly read rule records that are unrelated to the candidate intervention object.

[0229] In this embodiment, the rule index module first reads the intervention object in the candidate intervention record and the constraint category with written constraint value in the constraint set to form a rule index key, and then reads the target conflict rule set from the conflict rule base to make the candidate intervention record enter the trigger condition matching process corresponding to the intervention object and constraint category.

[0230] In a processing approach that does not employ closed-loop revalidation, adjustment requests generated from feedback data may directly override the original intervention items in the lifestyle intervention package.

[0231] In this embodiment, the closed-loop re-verification module first converts the intervention package adjustment request into an adjusted candidate intervention record, and then re-sends the adjusted candidate intervention record into the rule index module, conflict detection module, and disposal execution module for processing. After reprocessing and without writing a deletion mark, the intervention package generation module writes the corresponding intervention item into the lifestyle intervention package.

[0232] Based on embodiments 1-3, the working principle of this invention is as follows: The system first receives structured disease diagnosis information, medication information, comprehensive geriatric assessment information, and multi-source candidate lifestyle intervention measures; then, it converts the disease diagnosis information, medication information, and comprehensive geriatric assessment information into a constraint set, converts the multi-source candidate lifestyle intervention measures into a candidate intervention set, and writes a candidate intervention record for each candidate intervention measure; next, it establishes a rule index key based on the intervention object in the candidate intervention record and the constraint category with the constraint value already written in the constraint set, and reads the target conflict rule set from the conflict rule base; subsequently, it matches the structured field values ​​and the intervention parameters in the candidate intervention record with the trigger condition field to generate the target conflict rule hit result; when the hit result is empty, the treatment execution module writes a retention flag; when the hit result is not empty, the treatment execution module reads the treatment action field, treatment parameter field, and treatment priority field, and writes the execution condition, adjusts the execution parameters, writes the deletion flag, calls alternative measures, or writes the review flag for the candidate intervention record; the intervention package generation module writes the processed candidate intervention records without the deletion flag into the lifestyle intervention package. For candidate intervention records marked for deletion, the system invokes alternative candidate measures and generates alternative candidate intervention records. Then, it re-executes rule indexing and trigger condition matching on these alternative candidate intervention records. For adjusted candidate intervention records generated by closed-loop updates, the system re-executes rule indexing, trigger condition matching, and disposal writing processes before updating the lifestyle intervention package. Calling the target conflict rule set via the rule index key reduces traversal of irrelevant rules. Merging disposal parameters, verifying markers, and performing closed-loop re-verification reduces inconsistencies in disposal parameters caused by different modules handling the same candidate intervention record separately, and prevents adjusted candidate intervention records from bypassing the original constraint matching and disposal control processes.

[0233] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A personalized lifestyle intelligent intervention system for elderly people with comorbidities, characterized in that, It includes a data receiving module, a constraint set generation module, a candidate intervention set generation module, a rule indexing module, a conflict detection module, a treatment execution module, an intervention package generation module, and a closed-loop re-verification module; The data receiving module receives structured disease diagnosis information, medication information, comprehensive geriatric assessment information, and candidate lifestyle interventions. The constraint set generation module converts disease diagnosis information, medication information, and comprehensive geriatric assessment information into constraint sets. The candidate intervention set generation module receives dietary candidate interventions, exercise candidate interventions, sleep candidate interventions, psychological candidate interventions, social candidate interventions, and medication adherence candidate interventions, writes a candidate intervention record for each candidate intervention, and forms a candidate intervention set from multiple candidate intervention records; The rule indexing module constructs a rule index key based on the intervention object in the candidate intervention record and the constraint category in the constraint set, and retrieves the target conflict rule set from the conflict rule base based on the rule index key; The conflict detection module matches the intervention parameters in the candidate intervention record and the structured field values ​​in the constraint set with the trigger condition fields of each target conflict rule in the target conflict rule set to generate the target conflict rule hit result; When the result of the target conflict rule is empty, the handling execution module performs a retention mark write to the candidate intervention record; When the result of the target conflict rule hit is not empty, the handling execution module reads the handling action field, handling parameter field and handling priority field in the target conflict rule corresponding to the result of the target conflict rule hit, and writes the execution conditions, adjusts the execution parameters, writes the candidate measure deletion mark, and writes the alternative measure call or review mark according to the handling priority field. The intervention package generation module writes candidate intervention records that have been processed by the treatment execution module and have not been marked for deletion into the lifestyle intervention package; The closed-loop re-verification module receives compliance data, physiological monitoring feedback data, and abnormal event records corresponding to the lifestyle intervention package, generates an intervention package adjustment request, records the adjusted candidate intervention measures involved in the intervention package adjustment request as adjusted candidate intervention records, and re-sends the adjusted candidate intervention records into the rule index module, conflict detection module, and disposal execution module for processing before the lifestyle intervention package is updated.

2. The personalized lifestyle intelligent intervention system for elderly people with comorbidities according to claim 1, characterized in that: The constraint set includes a subset of disease constraints, a subset of medication constraints, and a subset of comprehensive geriatric assessment constraints; The constraint set generation module generates disease constraint subsets based on disease diagnosis information, medication constraint subsets based on medication information, and comprehensive geriatric assessment constraint subsets based on comprehensive geriatric assessment information. The constraint subset of the comprehensive assessment for the elderly includes functional status constraint records transformed from frailty status field, fall risk field, cognitive status field, nutritional risk field, daily living ability field, and care resource field.

3. The personalized lifestyle intelligent intervention system for elderly people with comorbidities according to claim 2, characterized in that: The constraint set generation module converts the frailty state field into frailty constraints, the fall risk field into fall constraints, the cognitive state field into execution complexity constraints, the nutritional risk field into dietary execution constraints, the daily living ability field into task execution constraints, and the care resource field into care assistance constraints, and writes the conversion results into the constraint subset of the comprehensive geriatric assessment.

4. The personalized lifestyle intelligent intervention system for elderly people with comorbidities according to claim 1, characterized in that: The candidate intervention record includes candidate source, intervention category, intervention target, execution frequency, execution duration, execution intensity, execution time, target parameters, intervention parameters, and integrity markers; Intervention categories include dietary intervention, exercise intervention, sleep intervention, psychological intervention, social intervention, and medication adherence intervention; When a candidate intervention set generation module is missing an intervention target, execution duration, execution intensity, or target parameter, it writes a field missing flag into the candidate intervention record.

5. A personalized lifestyle intelligent intervention system for elderly people with comorbidities according to claim 4, characterized in that: The conflict rule base called by the rule index module stores multiple conflicting rules; Each conflict rule includes a rule identifier, a rule version identifier, a rule source identifier, a constraint category field, a constraint object field, an intervention object field, a trigger condition field, a handling action field, a handling parameter field, and a handling priority field; The constraint category field stores the constraint category value that matches the constraint category in the constraint set; The constraint object field stores the constraint object value that matches the structured field value in the constraint set; The intervention object field stores the intervention object value that matches the intervention object in the candidate intervention record; The trigger condition field stores field matching conditions, field missing conditions, tag hit conditions, threshold comparison conditions, or intervention parameter out-of-bounds conditions; Actions include writing conditions to the disposal action field, parameter adjustment actions, writing candidate measure deletion markers, calling alternative measures, or writing review markers. The disposal parameter field stores the write parameters, adjustment parameters, replacement matching parameters, or review field parameters corresponding to the disposal action field. The priority field stores the execution sorting parameters when multiple conflicting rules for a target are hit simultaneously.

6. A personalized lifestyle intelligent intervention system for elderly people with comorbidities as described in claim 5, characterized in that: The rule indexing module creates rule index keys based on the constraint category field and the intervention object field; The rule indexing module reads intervention objects from candidate intervention records, reads constraint categories with written constraint values ​​from the constraint set, and uses the intervention object and the constraint category with written constraint values ​​to form a rule index key. It also reads the target conflict rule set from the conflict rule base.

7. A personalized lifestyle intelligent intervention system for elderly people with comorbidities as described in claim 6, characterized in that: The conflict detection module reads the constraint object field and trigger condition field of each target conflict rule in the target conflict rule set; The conflict detection module reads the structured field values ​​from the constraint set and the intervention parameters from the candidate intervention records; The conflict detection module performs equality matching on structured field values ​​and field matching conditions, matches empty fields with field missing conditions, matches structured field values ​​with tag hit conditions, compares structured field values ​​with threshold comparison conditions, and compares intervention parameters with upper and lower limits of intervention parameter out-of-bounds conditions. The conflict detection module generates a target conflict rule hit result when the structured field value corresponding to the constraint object field matches the trigger condition field, or when the intervention parameter in the candidate intervention record matches the trigger condition field.

8. A personalized lifestyle intelligent intervention system for elderly people with comorbidities according to claim 7, characterized in that: When the same candidate intervention record matches multiple target conflict rules, the handling execution module sorts the multiple target conflict rules according to the handling priority field; The disposal execution module writes execution conditions, adjusts execution parameters, writes candidate measure deletion markers, and writes alternative measure invocation or review markers for candidate intervention records according to the disposal action field and disposal parameter field; The treatment execution module will merge treatment parameters that can be written to the same candidate intervention record at the same time and write them into the candidate intervention record. When the action execution module writes the same parameter position in multiple action fields and the values ​​are inconsistent, or when the execution objects of multiple action fields are mutually exclusive, it writes a review mark to the candidate intervention record. When the disposal action field contains the action to write a candidate intervention deletion flag, the disposal execution module writes a deletion flag to the candidate intervention record to prevent it from being included in the lifestyle intervention package.

9. A personalized lifestyle intelligent intervention system for elderly people with comorbidities as described in claim 8, characterized in that: After a deletion mark is written to a candidate intervention record, the treatment execution module reads the intervention category and intervention target of the candidate intervention record with the deletion mark. The disposal execution module retrieves alternative candidate measures from the alternative measures library that have the same intervention category, the same intervention target, or meet the alternative matching parameters, and writes the alternative candidate measures into the candidate intervention set to form an alternative candidate intervention record; The rule indexing module re-invokes the set of rules for conflicting alternative objectives based on the intervention objects in the alternative candidate intervention records and the constraint categories in the constraint set; The conflict detection module will re-match the intervention parameters in the alternative candidate intervention record, the structured field values ​​in the constraint set, and the triggering conditions in the alternative target conflict rule set to generate the alternative target conflict rule hit result; When the result of the alternative target conflict rule is empty, the disposal execution module performs a retention mark write to the alternative candidate intervention record; When the result of the alternative target conflict rule is not empty, the disposal execution module reads the corresponding disposal action field, disposal parameter field, and disposal priority field, and writes the execution conditions, adjusts the execution parameters, writes the candidate measure deletion mark, or writes the review mark on the alternative candidate intervention record.

10. A personalized lifestyle intelligent intervention system for elderly people with comorbidities according to claim 1, characterized in that: The closed-loop re-verification module generates frequency adjustment requests, duration adjustment requests, or execution condition adjustment requests based on compliance data. The closed-loop re-verification module generates an execution parameter adjustment request based on physiological monitoring feedback data; The closed-loop re-verification module generates candidate measure deactivation requests, execution parameter downgrade requests, alternative candidate call requests, or review mark requests based on abnormal event records; The closed-loop re-verification module converts frequency adjustment requests, duration adjustment requests, execution condition adjustment requests, execution parameter adjustment requests, candidate measure deactivation requests, execution parameter downgrade requests, alternative candidate call requests, or review mark requests into adjusted candidate intervention measures, and records the adjusted candidate intervention measures as adjusted candidate intervention records; The rule indexing module re-invokes the target conflict rule set based on the intervention objects in the adjusted candidate intervention records and the constraint categories in the constraint set; The conflict detection module matches the intervention parameters in the adjusted candidate intervention record, the structured field values ​​in the constraint set, and the trigger conditions in the re-called target conflict rule set to generate the adjusted target conflict rule hit result. When the adjusted target conflict rule hit result is empty, the handling execution module performs a retention mark write to the adjusted candidate intervention record; When the result of the adjusted target conflict rule is not empty, the handling execution module reads the handling action field, handling parameter field and handling priority field in the adjusted target conflict rule, and writes the execution conditions, adjusts the execution parameters, writes the candidate measure deletion mark, and writes the alternative measure call or review mark on the adjusted candidate intervention record. The intervention package generation module writes the reprocessed candidate intervention records, which have not been marked for deletion, into the corresponding intervention items in the lifestyle intervention package.

Images