Hospitalization index and food image recognition-based blood lipid health management system

Abstract

The present application relates to the field of medical health information technology, and discloses a blood lipid health management system based on hospitalization indexes and food image recognition, which comprises a patient file grabbing module, a health risk reminding module, a personalized blood lipid target setting module, a food image recognition scanning module, a blood lipid index correlation calculation module, a health index analysis module and a blood lipid health visualization management module. Through deep integration of automatic grabbing of hospital diagnosis and treatment data and food image recognition AI technology outside the hospital, a continuous blood lipid risk dynamic monitoring and intervention closed loop is formed throughout the hospital, a three-dimensional quantitative evaluation model of daily blood lipid intake compliance Ge, blood lipid standard reaching prediction value Ys and blood vessel health comprehensive score Qc is constructed, dynamic correlation calculation of diet behavior, medication compliance and blood lipid biochemical indexes and blood vessel ultrasound structure indexes is realized, and decision basis for hierarchical intervention is provided for medical staff, so as to improve blood lipid standard reaching rate and patient self-management ability.

Description

Technical Field

[0001] This invention relates to the field of medical and health information technology, specifically to a blood lipid health management system based on hospitalization indicators and food image recognition. Background Technology

[0002] Currently, lipid management for high-risk groups and patients with cardiovascular and cerebrovascular diseases mainly relies on periodic outpatient follow-ups and patients' self-recorded lifestyles.

[0003] The existing management model has the following significant shortcomings: First, in-hospital medical data (such as inpatient medical records and laboratory test results) is fragmented from patients' daily life data outside the hospital (especially dietary intake), forming "information silos" that prevent doctors from obtaining a continuous and complete health profile for accurate assessment. Second, monitoring of diet, a key influencing factor, is extremely weak. Traditional reliance on manually recorded food diaries by patients suffers from significant subjective bias, difficulty in quantification, and low compliance, failing to provide objective and accurate data on the intake of nutrients such as cholesterol. Third, management is lagging and lacks foresight, typically intervening only after abnormal indicators are detected during follow-up examinations, missing the optimal adjustment window. Furthermore, there is a lack of intelligent tools to integrate and analyze blood indicators, vascular imaging, and behavioral data; management strategies are simplistic, patient participation and satisfaction are low, and the overall lipid target achievement rate and medical resource efficiency need improvement.

[0004] Therefore, there is an urgent need for an intelligent management system that can connect data from inside and outside the hospital, achieve precise quantification of diet, and provide dynamic early warning and personalized intervention. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a blood lipid health management system based on inpatient indicators and food image recognition. It has the advantages of breaking down data barriers, achieving precise quantitative management and intelligent early warning intervention, and building a collaborative management ecosystem. It solves the problems of data fragmentation, difficulty in dietary monitoring, lagging and passive management, insufficient doctor-patient collaboration, and lack of personalized full-cycle closed-loop management in the existing blood lipid management model.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a blood lipid health management system based on hospitalization indicators and food image recognition, including a patient file capture module, a health risk reminder module, a personalized blood lipid target setting module, a food image recognition and scanning module, a blood lipid indicator correlation calculation module, a health indicator analysis module, and a blood lipid health visualization management module;

[0007] The patient record capture module automatically captures all patient medical data, while also supporting medical staff to view, filter, and manually supplement special medical information. After the captured data enters the module, it automatically completes data classification, cleaning, and integration to form a personal health record. The health risk alert module receives personal health record information, identifies abnormal data, and has an alert function. When the patient's re-examination indicators show significant fluctuations, it automatically pushes an alert to the attending physician. The personalized lipid target setting module is operated by medical staff. Based on the patient's integrated diagnosis and treatment data, combined with the patient's ASCVD risk stratification, age, underlying diseases and medication, it sets personalized lipid control target values. The food image recognition and scanning module utilizes AI image recognition technology to allow patients to take pictures of their daily food intake via their mobile phones. It automatically identifies the types and amounts of food consumed and combines this with a blood lipid health food database to measure the daily intake of blood lipid-related components, providing patients with targeted dietary advice. The blood lipid index correlation calculation module calculates the daily blood lipid intake compliance Ge, the predicted blood lipid target Ys, and the comprehensive vascular health score Qc based on medical data, set blood lipid targets, and dietary intake data. This assists medical staff in developing personalized intervention plans and provides patients with intuitive references for intake and progress. The health indicator analysis module automatically identifies abnormal data correlations based on the calculation results and combined with the patient's dynamic blood lipid update data, and generates a personalized analysis report. The blood lipid health visualization management module gathers the data analysis results, solutions, strategies, and interactive commands from all modules and presents them centrally and interactively through a multi-terminal interface, enabling medical staff, patients, and their families to understand the patient's blood lipid health management status in real time.

[0008] Preferably, the patient record capture module includes an automatic data capture unit, a record dynamic update unit, and a data integration unit.

[0009] Preferably, the automatic data capture unit is deeply integrated with the hospital information system, electronic medical record system, and laboratory testing system through a secure interface to automatically capture all patient diagnosis and treatment data.

[0010] Preferably, the file dynamic update unit monitors and updates patient diagnosis and treatment data in real time, while supporting medical staff to view, filter, and manually supplement special diagnosis and treatment information; the data integration unit automatically classifies, cleans, and integrates the captured and updated data based on preset medical rules and data standards to form a unified personal health record.

[0011] Preferably, the blood lipid control target values ​​set by the personalized blood lipid target setting module include: low-density lipoprotein cholesterol target value, total cholesterol target value and triglyceride target value. At the same time, it can be linked to the vascular ultrasound examination results to comprehensively set the blood lipid target evaluation criteria. After the setting is completed, the target value is automatically synchronized to all subsequent modules.

[0012] Preferably, the blood lipid index correlation calculation module calculates the daily blood lipid intake compliance rate Ge, and the calculation formula is as follows: In the formula, Ge represents the compliance rate of daily blood lipid intake. This indicates the actual daily cholesterol intake. This indicates the recommended daily cholesterol intake. This represents the effect coefficient of medication use.

[0013] Preferably, the blood lipid index correlation calculation module calculates the predicted blood lipid target value Ys, and the calculation formula is as follows: In the formula, This indicates the patient's latest blood lipid level. This indicates the patient's average daily lipid intake compliance rate (Ge) over the past week. This represents the patient's medication adherence coefficient over the past week. This represents the conversion factor for the patient's average daily exercise volume over the past week. These represent the personalized influence weighting coefficients for diet, medication, and exercise, respectively.

[0014] Preferably, the visualization feedback module calculates a comprehensive vascular health score. The calculation formula is as follows: In the formula, This represents the personalized blood lipid control target value set for this patient. This indicates key quantitative indicators measured using the latest vascular ultrasound examination. This indicates the patient's personal baseline values ​​or clinically defined upper limits of normal for vascular ultrasound parameters. This represents the average daily lipid intake compliance rate (Ge) of the patient over the past 30 days. These represent the combined weighting coefficients for lipid target achievement, vascular structure improvement, and behavioral stability, respectively.

[0015] Preferably, the health indicator analysis module performs real-time risk assessment and stratified intervention decision support for the patient's blood lipid management status based on the calculation results of the daily blood lipid intake compliance rate Ge, the predicted blood lipid target value Ys, and the comprehensive vascular health score Qc, as follows: (1) Based on the daily blood lipid intake compliance rate Ge, it is used to judge the compliance of the patient's daily dietary cholesterol intake in real time and optimize the dietary guidance plan in a timely manner; (2) Based on the blood lipid target prediction value Ys, the probability of patients reaching their personalized control targets in the near future is predicted in real time. The prediction results are pushed out in stages. For patients with a low probability of reaching the target, medical staff are reminded to pay close attention, investigate problems in diet, medication, exercise and other aspects and intervene in a timely manner. (3) When the comprehensive vascular health score Qc > 80, maintain the current personalized management plan, continuously monitor changes in various indicators, and regularly send health consolidation suggestions; when 60 ≤ Qc ≤ 79, take measures to optimize the diet structure and strengthen exercise intervention, send targeted health guidance every week, and remind patients to complete a core indicator review once a month; when 40 ≤ Qc ≤ 59, take measures to immediately send medical and nursing intervention reminders and adjust the lipid-lowering drug plan, require patients to complete outpatient review within 1-2 weeks, increase the frequency of indicator monitoring, and closely track the improvement of vascular health status.

[0016] Preferably, the blood lipid health visualization management module serves as the system's visualization terminal, and its specific functions include: (1) Medical staff: View the patient's full health data and analysis reports, adjust blood lipid control targets, dietary and medication recommendations, receive patient inquiries, and push follow-up reminders; (2) Patient side: View personalized management plans, dietary advice and health data visualization results, upload food images, provide feedback on their own health status, and receive various reminders; (3) System side: Automatically archive patient health data, generate summary reports on blood lipid management, and support hospitals in calculating blood lipid target achievement rates and optimizing the allocation of medical resources.

[0017] Compared with existing technologies, this invention provides a blood lipid health management system based on hospitalization indicators and food image recognition, which has the following beneficial effects: 1. This invention constructs a three-dimensional quantitative assessment model that integrates daily blood lipid intake compliance Ge, blood lipid target prediction value Ys, and vascular health comprehensive score Qc. It dynamically correlates dietary behavior, medication adherence, exercise data, blood lipid biochemical indicators, and vascular ultrasound structural indicators to achieve precise quantitative assessment from microscopic dietary intake to macroscopic vascular health. This breakthrough overcomes the limitations of traditional single-point-of-time detection, enabling trend prediction and providing medical staff with a basis for tiered and graded intervention decisions.

[0018] 2. This invention establishes a closed-loop data system covering the entire cycle of inpatient treatment, outpatient monitoring, intervention and adjustment, and follow-up evaluation by deeply integrating inpatient treatment data automatic capture technology with outpatient food image recognition AI technology. This breaks down information barriers between inpatient and outpatient settings to achieve continuity of medical behavior. Furthermore, it objectively records dietary intake through visual recognition technology to reduce subjective recall bias and constructs a real-time two-way feedback mechanism, enabling the system to achieve the beneficial effect of improving patient self-management compliance.

[0019] 3. This invention uses machine learning algorithms to personalize the weight coefficients of diet, medication and exercise, and dynamically adjusts the weights of blood lipid control targets and vascular health scores in conjunction with ASCVD risk stratification. This enables the system to shift from standardized population management to individualized precision management. Furthermore, the accuracy of the prediction model is continuously optimized as patient data accumulates, ultimately enabling the system to automatically identify abnormal data correlations and reveal potential risk factors. Attached Figure Description

[0020] Figure 1 This is a system flowchart of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Please see Figure 1 The blood lipid health management system based on hospitalization indicators and food image recognition includes a patient file capture module, a health risk reminder module, a personalized blood lipid target setting module, a food image recognition and scanning module, a blood lipid indicator correlation calculation module, a health indicator analysis module, and a blood lipid health visualization management module. The patient record capture module automatically captures all patient medical data, while also allowing medical staff to view, filter, and manually supplement special medical information. After the captured data enters the module, it automatically completes data classification, cleaning, and integration to form a personal health record. The health risk alert module receives personal health record information, identifies abnormal data, and has an alert function. When a patient's re-examination indicators (such as blood lipids, liver and kidney function) show significant fluctuations, it automatically pushes an alert to the attending physician. The personalized lipid target setting module is operated by medical staff (attending physicians and health managers). Based on the patient's integrated diagnosis and treatment data, combined with the patient's ASCVD (atherosclerotic cardiovascular disease) risk stratification, age, underlying diseases and medication, a personalized lipid control target value is set. After the setting is completed, the target value is automatically synchronized to all subsequent analysis and visualization modules, serving as the core benchmark for lipid management and ensuring that the target is accurately adapted to each patient. As the core of outpatient lipid management, the food image recognition scanning module deeply integrates AI image recognition technology, allowing patients to take pictures of their daily food intake (before or after meals) using their mobile phones. The module automatically identifies the types and amounts of food consumed, and combines them with a lipid-healthy food database to accurately measure the daily intake of lipid-related components (cholesterol, triglycerides). The module has an error correction function, allowing patients to manually adjust recognition deviations. Medical staff can view the patient's daily dietary intake records and, in conjunction with lipid targets, provide targeted dietary advice to patients, achieving real-time correlation between dietary intake, lipid components, and targets. The lipid index correlation calculation module is the core processing module of the system. It receives the diagnosis and treatment data from the previous patient file capture module, the lipid targets from the personalized lipid target setting module, and the dietary intake data from the food image recognition and scanning module. By calculating the daily lipid intake compliance Ge, the predicted value of lipid target achievement Ys, and the comprehensive vascular health score Qc, it realizes the dynamic correlation analysis of dietary intake, exercise behavior, medication adherence and lipid index, assists medical staff in formulating personalized intervention plans, and provides patients with intuitive reference for intake and progress. The health indicator analysis module integrates multi-dimensional health data of patients based on the calculation results of daily blood lipid intake compliance (Ge), predicted blood lipid levels (Ys), and vascular health comprehensive score (Qc). This includes blood indicators (blood lipids, liver and kidney function, etc.), vascular ultrasound results, dietary intake data, medication data, and wearable device data. The module performs dynamic analysis, allowing medical staff to view patients' historical data trends (such as changes in blood lipid indicators and dietary intake over the past month), automatically identify abnormal data correlations (such as low dietary suitability and elevated LDL-C), generate personalized analysis reports, and clarify current problems in blood lipid management (such as excessive diet or irregular medication). The blood lipid health visualization management module serves as the system's visualization terminal, aggregating and integrating data analysis results, strategies, and interactive commands from all preceding modules. It presents and interacts centrally through a multi-terminal interface, allowing patients and their families to view their health data at any time and intuitively grasp the progress of blood lipid management. Relevant visualization information is synchronized to medical staff, enabling them to understand the patient's outpatient management status in real time. This achieves two-way feedback between medical staff monitoring and patient self-checking. This mechanism continuously drives the execution and dynamic iteration of personalized management plans, supporting users in proactively and continuously managing their blood lipid health.

[0023] The patient record capture module serves as the system's data foundation and is divided into an automatic data capture unit, a dynamic record update unit, and a data integration unit based on the data processing flow and core functions.

[0024] The automatic data capture unit integrates deeply with the hospital information system (HIS), electronic medical record system (EMR), and laboratory testing system through secure and compliant interfaces. It automatically captures all patient diagnosis and treatment data without manual entry, ensuring data authenticity, real-time performance, and accuracy. The specific content captured includes: inpatient medical records (procedure notes, medication records, nursing records), outpatient follow-up records (complete blood lipid panel, liver and kidney function tests, blood glucose, and other blood indicators), and vascular imaging reports (carotid ultrasound, etc.).

[0025] The dynamic record update unit monitors and updates patient treatment data in real time, while also allowing medical staff to view, filter, and manually supplement special treatment information to improve patient health records and ensure the integrity of treatment data. The data integration unit automatically classifies, cleans, and integrates the captured and updated data based on preset medical rules and data standards to form a unified personal health record, providing data support for subsequent personalized management.

[0026] The personalized lipid target setting module allows for the setting of lipid control target values, including low-density lipoprotein cholesterol (LDL-C) target values, total cholesterol (TC) target values, and triglyceride (TG) target values. It can also be linked to vascular ultrasound examination results to comprehensively set the criteria for achieving lipid targets. After the settings are completed, the target values ​​are automatically synchronized to all subsequent analysis and visualization modules, serving as the core benchmark for lipid management and ensuring that the targets are accurately adapted to each patient.

[0027] The blood lipid index correlation calculation module calculates the daily blood lipid intake compliance rate Ge, and the calculation formula is as follows: In the formula, Ge represents the compliance of daily blood lipid intake, reflecting the degree to which the patient's daily cholesterol intake matches the personalized recommendation. This represents the actual daily cholesterol intake, which is the total daily cholesterol intake of the patient as measured by the food image recognition scanning module (unit: mg). This indicates the recommended daily cholesterol intake, which is set based on the patient's individualized blood lipid target and underlying diseases (e.g., ≤200mg per day for high-risk individuals). This represents the drug effect coefficient, which is set according to the type and dosage of lipid-lowering drugs taken by the patient (the value ranges from 0 to 5, and the stronger the lipid-lowering effect of the drug, the smaller the coefficient).

[0028] The advantages are: by calculating the daily lipid intake compliance rate (Ge), it achieves precise quantitative assessment of dietary behavior, transforming the abstract concept of healthy eating into intuitive numerical feedback, thereby improving patients' adherence to dietary self-management; and by establishing a drug-diet interaction correction mechanism, it uses the drug effect coefficient. By dynamically adjusting compliance thresholds, we can avoid excessive restrictions on patients taking medication through simple dietary control. This allows nutritionists to quickly identify high-risk eating days based on Ge values ​​and promptly push intervention suggestions, ultimately achieving early detection and early correction of dietary problems.

[0029] The blood lipid index correlation calculation module calculates the predicted value Ys for achieving blood lipid targets, and the calculation formula is as follows: In the formula, This indicates the patient's latest blood lipid level (such as low-density lipoprotein cholesterol LDL-C, unit: millimoles per liter, mmol / L), and the data comes from the patient record capture module; Ge represents the patient's average daily lipid intake compliance over the past week (7 days), which reflects short-term dietary adherence. This represents the patient's medication adherence coefficient over the past week, with a value ranging from 0 to 1. 1 indicates that the patient has fully followed the doctor's instructions to take the medication, while 0 indicates that the patient has not taken the medication. The data is obtained from the patient's medication attendance record or pharmacy dispensing record. The conversion factor (unit: metabolic equivalent - minute, MET-min) represents the patient's average daily exercise volume over the past week. The data is collected and calculated by wearable devices (such as health bracelets). These represent the personalized influence weighting coefficients of diet, medication, and exercise, respectively. These three coefficients are set in the initial stage of the system based on the evidence-based medicine model of the patient population, and are then personalized and calibrated based on the patient's long-term data through machine learning to improve prediction accuracy.

[0030] The advantages are: by calculating the predicted value Ys for achieving target blood lipid levels, a multi-dimensional prediction model combining current indicators and recent behaviors is constructed, enabling the system to predict blood lipid control trends 1-2 weeks in advance, thus providing an intervention window for both doctors and patients; by integrating the weighted proportions of three major intervention elements—diet, medication, and exercise—(… Machine learning continuously optimizes individualized prediction parameters, enabling the prediction model to become increasingly accurate as patient data accumulates. This establishes a visualized target achievement probability assessment system, transforming complex medical indicators into intuitive target achievement likelihoods. This lowers the barrier to understanding for patients while also enhancing their confidence and motivation in treatment.

[0031] The visualization feedback module calculates the comprehensive vascular health score Qc, and the calculation formula is as follows: In the formula, This represents the personalized lipid control target value (such as LDL-C target, unit: mmol / L) set for this patient, and its value is set by the personalized lipid target setting module; This indicates key quantitative indicators measured using the latest vascular ultrasound examinations (such as carotid ultrasound), such as maximum plaque thickness (unit: mm) or luminal stenosis rate (percentage, %), with data sourced from the patient record capture module; This indicates the patient's individual baseline or clinically defined upper limit of normal vascular ultrasound parameters, used to measure the relative degree of changes in vascular structure; This represents the average daily lipid intake compliance (Ge) of patients over the past 30 days (one month), reflecting the stability and adherence of dietary behavior in the medium term. These represent the combined weighting coefficients for lipid target achievement, vascular structure improvement, and behavioral stability, with the weights varying according to different management stages (e.g., greater emphasis during the acute phase). More attention is paid to the long-term maintenance period. (Or the specific circumstances of the patient can be adjusted by medical staff.)

[0032] The advantages are: by calculating the comprehensive vascular health score Qc, it integrates three dimensions—biochemical indicators (blood lipids), structural indicators (vascular ultrasound), and behavioral indicators (long-term dietary adherence)—to construct a comprehensive health profile, effectively avoiding the one-sidedness of single-indicator assessment; and by introducing a time-dimensional hierarchical weighting design (…). It can dynamically adjust according to the acute / maintenance phase, enabling the system's scoring system to automatically adapt to management priorities at different stages of the disease, thus improving the clinical applicability of the assessment; by establishing a medium- to long-term health trend tracking mechanism, the average compliance rate over 30 days is tracked. It can reflect the stability of patient behavior, thereby effectively identifying the difference between occasional violations and persistent adverse events, and ultimately providing a quantitative basis for medical staff to develop differentiated follow-up strategies.

[0033] The health indicator analysis module, based on the calculation results of daily lipid intake compliance (Ge), predicted lipid target achievement (Ys), and comprehensive vascular health score (Qc), provides real-time analysis, risk assessment, and stratified intervention decision support for patients' lipid management status, as detailed below: (1) Based on the daily blood lipid intake compliance rate Ge, it is used to judge the compliance of the patient's daily dietary cholesterol intake in real time, and push targeted dietary adjustment suggestions at the same time. The compliance data is also synchronized to the medical staff, so that the medical staff can dynamically grasp the patient's dietary implementation and optimize the dietary guidance plan in a timely manner. (2) Based on the blood lipid target prediction value Ys, the probability of patients reaching personalized control targets in the near future (1-2 weeks) is predicted in real time, and reminders are pushed in stages according to the prediction results. For patients with a low probability of reaching the target, medical staff are reminded to pay close attention, investigate problems in diet, medication, exercise and other aspects and intervene in a timely manner. (3) When the comprehensive vascular health score Qc > 80 (excellent), measures are taken to maintain the current personalized management plan, continuously monitor changes in various indicators, and regularly push health consolidation suggestions; when 60 ≤ Qc ≤ 79 (good), measures are taken to optimize the diet structure and strengthen exercise intervention, targeted health guidance is pushed weekly, and patients are reminded to complete a core indicator re-examination once a month; when 40 ≤ Qc ≤ 59 (average), measures are taken to immediately push medical and nursing intervention reminders and adjust the lipid-lowering drug plan (to be implemented after evaluation by medical and nursing staff), requiring patients to complete outpatient re-examination within 1-2 weeks, and simultaneously increasing the frequency of indicator monitoring to closely track the improvement of vascular health status; Meanwhile, this module allows medical staff to view patients' historical data trends (such as changes in blood lipid levels, changes in dietary intake, and fluctuation trends of Ge, Ys, and Qc values ​​over the past month), automatically identify abnormal data correlations (such as low dietary compliance with elevated LDL-C, and decreased Qc accompanied by decreased Ys), generate personalized analysis reports, clarify the core issues in current blood lipid management, and provide solid data support for subsequent precise intervention.

[0034] The advantages are: by establishing a closed-loop management chain of real-time monitoring, risk warning, and tiered intervention, and based on the three core indicators of Ge, Ys, and Qc, it achieves multi-level analysis from micro-behavior (daily diet) to macro-outcome (vascular structure), ensuring that intervention measures are accurately matched with the patient's current health status. The system can automatically identify abnormal combinations of cross-indicators (such as the time-lag correlation between decreased dietary compliance and increased LDL-C, and the correlation between insufficient exercise and stagnant vascular scores), thereby revealing potential risk factors that are difficult to discover with traditional single-indicator analysis. At the same time, it supports one-click generation of structured analysis reports by medical staff, transforming complex multidimensional data into clinical decision support documents that include problem diagnosis, trend prediction, and intervention suggestions, ultimately greatly improving the efficiency of the system's chronic disease management.

[0035] The blood lipid health visualization management module serves as the system's visualization terminal, and its specific functions include: (1) Medical staff: View the patient's full health data and analysis reports, adjust blood lipid control targets, dietary and medication recommendations, receive patient inquiries, and push follow-up reminders; (2) Patient side: View personalized management plans, dietary advice, and health data visualization results; upload food images; provide feedback on their own health status; and receive various reminders. (3) System side: Automatically archive patient health data, generate blood lipid management summary reports, support hospital-level statistics on blood lipid target achievement rate, optimize medical resource allocation, and build a full-cycle blood lipid health management system of "in-hospital diagnosis and treatment - out-of-hospital monitoring - intervention and adjustment - re-examination and evaluation".

[0036] The advantages are: by constructing a digital management ecosystem that integrates medical staff, patients, and the system, and by achieving real-time synchronization of information across the three terminals through a unified data platform, the data barriers between the hospital and other institutions are broken down, thereby ensuring the continuity and consistency of medical practices. Specifically, by providing full-lifecycle blood lipid health record management, it supports seamless transition from acute inpatient treatment to stable home rehabilitation; through automatic archiving and summary reporting functions, it forms a traceable and assessable management loop, enabling the system to optimize the allocation of medical resources; and through hospital-level blood lipid target achievement rate statistics and automatic marking of high-risk patients, it helps management identify management weaknesses, rationally allocate follow-up resources for nutritionists and cardiologists, and ultimately improve the group benefits and cost-effectiveness ratio of the system's chronic disease management.

[0037] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A blood lipid health management system based on hospitalization indicators and food image recognition, characterized in that, It includes a patient record capture module, a health risk reminder module, a personalized blood lipid target setting module, a food image recognition and scanning module, a blood lipid index correlation calculation module, a health index analysis module, and a blood lipid health visualization management module; The patient record capture module automatically captures all patient medical data, while also supporting medical staff to view, filter, and manually supplement special medical information. After the captured data enters the module, it automatically completes data classification, cleaning, and integration to form a personal health record. The health risk alert module receives personal health record information, identifies abnormal data, and has an alert function. When the patient's re-examination indicators show significant fluctuations, it automatically pushes an alert to the attending physician. The personalized lipid target setting module is operated by medical staff. Based on the patient's integrated diagnosis and treatment data, combined with the patient's ASCVD risk stratification, age, underlying diseases and medication, it sets personalized lipid control target values. The food image recognition and scanning module utilizes AI image recognition technology to allow patients to take pictures of their daily food intake via their mobile phones. It automatically identifies the types and amounts of food consumed and combines this with a blood lipid health food database to measure the daily intake of blood lipid-related components, providing patients with targeted dietary advice. The blood lipid index correlation calculation module calculates the daily blood lipid intake compliance Ge, the predicted blood lipid target Ys, and the comprehensive vascular health score Qc based on medical data, set blood lipid targets, and dietary intake data. This assists medical staff in developing personalized intervention plans and provides patients with intuitive references for intake and progress. The health indicator analysis module automatically identifies abnormal data correlations based on the calculation results and combined with the patient's dynamic blood lipid update data, and generates a personalized analysis report. The blood lipid health visualization management module gathers the data analysis results, solutions, strategies, and interactive commands from all modules and presents them centrally and interactively through a multi-terminal interface, enabling medical staff, patients, and their families to understand the patient's blood lipid health management status in real time.

2. The blood lipid health management system based on hospitalization indicators and food image recognition according to claim 1, characterized in that: The patient record retrieval module includes an automatic data retrieval unit, a dynamic record update unit, and a data integration unit.

3. The blood lipid health management system based on hospitalization indicators and food image recognition according to claim 2, characterized in that: The automatic data capture unit is deeply integrated with the hospital information system, electronic medical record system, and laboratory testing system through a secure interface, automatically capturing all patient diagnosis and treatment data.

4. The blood lipid health management system based on hospitalization indicators and food image recognition according to claim 2, characterized in that: The dynamic record update unit monitors and updates patient diagnosis and treatment data in real time, while also supporting medical staff to view, filter, and manually supplement special diagnosis and treatment information; the data integration unit automatically classifies, cleans, and integrates the captured and updated data based on preset medical rules and data standards to form a unified personal health record.

5. The blood lipid health management system based on hospitalization indicators and food image recognition according to claim 1, characterized in that: The personalized blood lipid target setting module sets blood lipid control target values ​​including: low-density lipoprotein cholesterol target value, total cholesterol target value and triglyceride target value. It can also be linked to vascular ultrasound examination results to comprehensively set blood lipid target evaluation criteria. After the setting is completed, the target value is automatically synchronized to all subsequent modules.

6. The blood lipid health management system based on hospitalization indicators and food image recognition according to claim 1, characterized in that: The blood lipid index correlation calculation module calculates the daily blood lipid intake compliance rate Ge, and the calculation formula is as follows: In the formula, Ge represents the compliance rate of daily blood lipid intake. This indicates the actual daily cholesterol intake. This indicates the recommended daily cholesterol intake. This represents the effect coefficient of medication use.

7. The blood lipid health management system based on hospitalization indicators and food image recognition according to claim 1, characterized in that: The blood lipid index correlation calculation module calculates the predicted blood lipid target value Ys, and the calculation formula is as follows: In the formula, This indicates the patient's latest blood lipid level. This indicates the patient's average daily lipid intake compliance rate (Ge) over the past week. This represents the patient's medication adherence coefficient over the past week. This represents the conversion factor for the patient's average daily exercise volume over the past week. These represent the personalized influence weighting coefficients for diet, medication, and exercise, respectively.

8. The blood lipid health management system based on hospitalization indicators and food image recognition according to claim 1, characterized in that: The visualization feedback module calculates the comprehensive vascular health score Qc, and the calculation formula is as follows: In the formula, This represents the personalized blood lipid control target value set for this patient. This indicates key quantitative indicators measured using the latest vascular ultrasound examination. This indicates the patient's personal baseline values ​​or clinically defined upper limits of normal for vascular ultrasound parameters. This represents the average daily lipid intake compliance rate (Ge) of the patient over the past 30 days. These represent the combined weighting coefficients for lipid target achievement, vascular structure improvement, and behavioral stability, respectively.

9. The blood lipid health management system based on hospitalization indicators and food image recognition according to claim 1, characterized in that: The health indicator analysis module performs real-time risk assessment and stratified intervention decision support for patients' blood lipid management status based on the calculation results of daily blood lipid intake compliance (Ge), blood lipid target prediction value (Ys), and vascular health comprehensive score (Qc), as follows: (1) Based on the daily blood lipid intake compliance rate Ge, it is used to judge the compliance of the patient's daily dietary cholesterol intake in real time and optimize the dietary guidance plan in a timely manner; (2) Based on the blood lipid target prediction value Ys, the probability of patients reaching their personalized control targets in the near future is predicted in real time. The prediction results are pushed out in stages. For patients with a low probability of reaching the target, medical staff are reminded to pay close attention, investigate problems in diet, medication, exercise and other aspects and intervene in a timely manner. (3) When the comprehensive vascular health score Qc > 80, maintain the current personalized management plan, continuously monitor changes in various indicators, and regularly send health consolidation suggestions; when 60 ≤ Qc ≤ 79, take measures to optimize the diet structure and strengthen exercise intervention, send targeted health guidance every week, and remind patients to complete a core indicator review once a month; when 40 ≤ Qc ≤ 59, take measures to immediately send medical and nursing intervention reminders and adjust the lipid-lowering drug plan, require patients to complete outpatient review within 1-2 weeks, increase the frequency of indicator monitoring, and closely track the improvement of vascular health status.

10. The blood lipid health management system based on hospitalization indicators and food image recognition according to claim 1, characterized in that: The blood lipid health visualization management module serves as the system's visualization terminal, and its specific functions include: (1) Medical staff: View the patient's full health data and analysis reports, adjust blood lipid control targets, dietary and medication recommendations, receive patient inquiries, and push follow-up reminders; (2) Patient side: View personalized management plans, dietary advice and health data visualization results, upload food images, provide feedback on their own health status, and receive various reminders; (3) System side: Automatically archive patient health data, generate summary reports on blood lipid management, and support hospitals in calculating blood lipid target achievement rates and optimizing the allocation of medical resources.

Images