Nutrition matching management system for stroke dysphagia patient

Abstract

The present application relates to the technical field of meal management, and provides a kind of nutrition meal management system for stroke dysphagia patient, including microprocessor and multiple function modules.Patients information acquisition module collects the basic information of patient;Diet record module records the diet of patient and calculates food intolerance index;Symptom monitoring module tracks the symptom change of patient and assesses symptom severity;Nutrition requirement analysis module customizes the nutrition requirement of patient and monitors actual intake;Meal compliance evaluation module monitors the diet execution of patient and evaluates compliance;Meal adaptability evaluation module evaluates the adaptability of patient to current meal plan according to the above index;Data storage module stores the data collected and generated by each module;Diet behavior pattern recognition module identifies the long-term diet behavior pattern of patient using long short-term memory network algorithm;Each module is connected with each other, and works cooperatively, to form a complete meal management system.

Description

Technical Field

[0001] This invention relates to the field of meal management technology, and more specifically, to a nutritional meal management system for stroke patients with dysphagia. Background Technology

[0002] Stroke is an acute cerebrovascular disease with a high disability rate. Many stroke patients experience dysphagia, a condition characterized by difficulty and discomfort during eating and swallowing. This symptom manifests as difficulty swallowing, choking, aspiration, and even weight loss, severely impacting the patient's nutritional intake and quality of life. If dysphagia is not effectively managed, it can lead to malnutrition, pneumonia, and other complications, further aggravating the patient's condition. Therefore, scientific nutritional management and personalized dietary plans for stroke patients with dysphagia have significant clinical value in the rehabilitation process.

[0003] The importance of nutritional management in stroke patients with dysphagia is widely recognized. For example, existing published literature (Frequent and personalized nutritional support leads to improved nutritional status, activities of daily living, and dysphagia after stroke, Nutrition, 2020, Shimazu S. et al.) studied the positive effects of personalized meal monitoring and management on stroke patients with dysphagia. The study showed that precise meal monitoring, including regular adjustments to diet texture, energy, and protein intake, significantly improved patients' swallowing function and self-care ability. The literature emphasizes that frequent and personalized meal management, monitoring patients' daily food intake, and adjusting meal content in a timely manner to meet their recovery needs are crucial. This meal management approach not only plays a positive role in improving patients' nutritional status but also reduces the risk of aspiration and malnutrition by monitoring swallowing responses and food intake, providing important support for stroke patients in the recovery period.

[0004] Existing meal management systems provide patients with personalized dietary recommendations, typically including functions such as dietary records, symptom monitoring, nutritional analysis, and adherence assessment. For example, Reference 2...

[0005] In their paper "Intelligent System for a Personalized Diet of Obese Patients with Cancer" (IEEE International Conference on Electrical and Power Engineering, 2014), a research team developed an intelligent system designed to provide personalized meal management for cancer patients. This system integrates data collection, nutritional needs analysis, and dietary monitoring functions. It can monitor patients' nutritional intake, manage their meal arrangements, and adjust dietary plans in real time based on their health data. By combining patients' daily dietary records and health indicators, the system provides personalized meal management plans, helping to improve patient adherence and overall health.

[0006] However, the meal management system in Reference 2 still has some limitations. For example, the system mainly relies on dietary and symptom data entered by the patient, which may lead to inaccurate or incomplete data. Furthermore, the system has limited capabilities in assessing meal fit and cannot fully capture the complex relationship between fluctuations in the patient's condition and dietary intake, particularly in terms of multi-dimensional data integration and intelligent analysis. These issues limit the system's potential for application in broader clinical scenarios.

[0007] To address the aforementioned problems, a technical solution is provided. Summary of the Invention

[0008] To overcome the aforementioned deficiencies of existing technologies, this invention provides a nutritional meal management system for stroke patients with dysphagia. By integrating modules for patient information collection, dietary records, symptom monitoring, nutritional needs analysis, meal compliance assessment, meal adaptation assessment, data storage, and dietary behavior pattern recognition, and utilizing sensor technology, image recognition, mathematical models, and machine learning algorithms, it achieves comprehensive, accurate, and intelligent management of the patient's dietary status. This solves the problems of inaccurate dietary records, imprecise symptom correlation analysis, non-dynamic nutritional needs analysis, unsystematic meal assessment, and lack of long-term dietary behavior pattern recognition in existing technologies.

[0009] To achieve the above objectives, the present invention provides the following technical solution:

[0010] A nutritional meal management system for stroke patients with dysphagia includes a patient information collection module, a diet record module, a symptom monitoring module, a nutritional needs analysis module, a meal compliance assessment module, a meal adaptation assessment module, a data storage module, and a dietary behavior pattern recognition module. The patient information collection module is connected to the data storage module and the nutritional needs analysis module; the diet record module is connected to the data storage module, the symptom monitoring module, and the meal adaptation assessment module; the symptom monitoring module is connected to the data storage module and the meal adaptation assessment module; the nutritional needs analysis module is connected to the data storage module and the meal adaptation assessment module; and the meal compliance assessment module is connected to the data storage module and the meal behavior pattern recognition module. The meal adaptation assessment module is connected to the data storage module; the dietary behavior pattern recognition module is connected to the data storage module, the meal adaptation assessment module, the nutritional requirements analysis module, the meal compliance assessment module, and the dietary record module; the meal adaptation assessment module is used to construct a meal adaptation assessment model based on the patient's nutritional intake index, symptom severity index, patient compliance index, and food intolerance index transmitted by the dietary record module, symptom monitoring module, nutritional requirements analysis module, and meal compliance assessment module, to assess the adaptation index of stroke patients with dysphagia to the current meal plan, wherein the formula of the meal adaptation assessment model is:

[0011]

[0012] In the formula, ADI is the meal fit index, NII is the nutritional intake index, SSI is the symptom severity index, PCI is the patient compliance index, and FII is the food intolerance index.

[0013] As a further embodiment of the present invention, the diet recording module includes a diet data acquisition unit, a symptom association recording unit, and a food intolerance index calculation unit. The diet data acquisition unit measures food weight using a weight sensor and uses image recognition technology to capture photos of the meal preparation to identify and record food types. The diet data acquisition unit comprehensively utilizes weight data and image recognition results to record the types and weights of food ingested by the patient. The symptom association recording unit records the patient's daily symptoms, including symptom type, occurrence time, duration, and severity level. The food intolerance index calculation unit uses the data provided by the diet data acquisition unit and the symptom association recording unit to assess the patient's food intolerance index, including the following steps:

[0014] Step A1 involves analyzing the relationship between food intake weight and symptom presentation within a specific time window prior to symptom onset to quantify the association strength between food and symptoms. For each food and symptom, a food-symptom association matrix is ​​established, and its association index is calculated. The formula for calculating the association index between food and symptoms is as follows:

[0015]

[0016] In the formula, R fs Y represents the correlation index between food f and symptom s. s (τ) represents the severity rating of symptom s at time τ, where τ is the time point in time when the symptoms were recorded, and X f (τ-δ) represents the weight of food f ingested δ time before the onset of symptoms;

[0017] Step A2: Based on the food-symptom association matrix, construct a food intolerance index assessment model to evaluate the patient's degree of intolerance to the prepared food. The formula for the food intolerance index assessment model is as follows:

[0018]

[0019] In the formula, FII is the food intolerance index, F is the number of food types in the meal plan, S is the number of symptoms considered, and W is the number of symptoms considered. f V represents the weight of food f in the overall meal preparation, determined based on the food weight. s R represents the weight of symptom s. fs The correlation index between food f and symptom s.

[0020] As a further embodiment of the present invention, the symptom monitoring module includes a symptom detection unit and a symptom feature extraction unit. The symptom monitoring unit is used to track and record changes in the patient's symptoms, including dysphagia, choking, choking while drinking, and weight changes, and transmits these symptoms to the symptom feature extraction unit. The symptom feature extraction unit is used to extract features based on the data transmitted by the symptom monitoring unit, obtaining a dysphagia index, choking frequency, choking frequency while drinking, and weight change percentage, and constructing a symptom severity assessment model to obtain a symptom severity index. The formula for the symptom severity assessment model is:

[0021]

[0022] In the formula, SSI is the symptom severity index, T is the monitoring period, AP(t) is the dysphagia index at time t, DF(t) is the choking frequency at time t, BL(t) is the choking frequency when drinking water at time t, and WC(t) is the percentage change in weight at time t.

[0023] As a further embodiment of the present invention, the nutritional requirements analysis module includes a nutritional requirements customization unit, a nutritional intake monitoring unit, and a nutritional requirements analysis unit; the nutritional requirements customization unit is used to record the patient's daily recommended nutrient intake; the nutritional intake monitoring unit is used to monitor the patient's actual nutrient intake at each meal using a food nutrient composition monitoring and analysis instrument; the nutrients include calories, protein, fat, carbohydrates, vitamins, and minerals; the nutritional requirements analysis unit is used to construct a nutritional intake index assessment model by comparing the patient's actual nutrient intake with their recommended nutrient intake, and to assess the patient's nutritional intake index, including the following steps:

[0024] Step B1: Obtain the patient's recommended daily nutrient intake from the nutrition needs customization unit;

[0025] Step B2: Obtain the patient's actual daily nutritional intake from the nutritional intake monitoring unit;

[0026] Step B3: Construct a nutritional intake index assessment model to obtain the patient's nutritional intake index. The formula for the nutritional intake index assessment model is as follows:

[0027]

[0028] In the formula, NII is the patient's nutritional intake index, n is the quantity of nutrients, and w i Let be the weight of the i-th nutrient component, where min() is the min function, A i T represents the actual daily nutrient intake. i This refers to the recommended daily intake of nutrients.

[0029] As a further aspect of the present invention, the meal adherence assessment module includes a diet execution monitoring unit and a adherence index assessment unit. The diet execution monitoring unit records the number of meals the patient actually eats each day and the start and end times of each meal. It also measures the weight of food actually ingested by the patient at each meal using a weight sensor and, combined with a food nutrition database, obtains the average daily calorie intake. The adherence index assessment unit compares the number of meals the patient actually eats with the planned number of meals in the meal plan to assess adherence to meal frequency; compares the average calorie intake of the patient with the planned average calorie intake to analyze the implementation of calorie control; compares the average meal duration recorded by the patient with the planned average meal duration; and integrates these parameters into a quantitative adherence index to construct a patient adherence index assessment model, obtaining the patient's meal adherence index. The formula for the adherence index assessment model is:

[0030]

[0031] In the formula, PCI is the patient's meal compliance index, ME is the number of meals actually performed, MP is the number of planned meals, AE is the average actual calorie intake, AP is the average planned calorie intake, TE is the average actual recorded meal duration, and TP is the average planned meal duration.

[0032] As a further aspect of the present invention, the dietary behavior pattern recognition module is used to identify the patient's long-term dietary behavior pattern using a long short-term memory network algorithm, including the following steps:

[0033] Step S1: Construct a dataset for identifying dietary behavior patterns. The dataset includes a meal fit index, a nutrient intake index, a patient compliance index, and a food intolerance index. The meal fit index is calculated by the meal fit assessment module; the nutrient intake index is calculated by the nutrient requirements analysis module; the patient compliance index is calculated by the meal compliance assessment module; and the food intolerance index is calculated by the dietary record module. The dataset covers the patient's daily data over the past 6 months.

[0034] Step S2 involves preprocessing the dataset, including organizing the daily data into a time series format, standardizing each index, constructing a sliding window with a window size of 30 days and a step size of 1 day, and dividing the preprocessed dataset into training, validation, and test sets in an 8:1:1 ratio.

[0035] Step S3: Construct a long short-term memory network model, which includes an input layer, a long short-term memory network layer, a discard layer, and a fully connected layer; the input layer receives meal fit index, nutrient intake index, patient compliance index, and food intolerance index.

[0036] Step S4: Train the Long Short-Term Memory network model using the training set data, and monitor the model performance on the validation set using the Adam optimizer and early stopping strategy.

[0037] Step S5: Evaluate the model performance on the test set, calculate the root mean square error and mean absolute error for each index, and assess the model's prediction accuracy; use the trained model to predict the index for the next 30 days, analyze the prediction results, and output the identified patient dietary behavior pattern type.

[0038] As a further aspect of the present invention, the patient information collection module is used to collect basic patient information, including age, gender, height, weight, disease stage, comorbidities, and allergy history; the data storage module is used to store data from various modules, including basic information collected by the patient information collection module, food types, weights, and food intolerance index recorded by the diet record module; dysphagia index, choking frequency, choking frequency when drinking water, percentage weight change, and symptom severity index obtained by the symptom monitoring module; nutrient intake and nutrient intake index obtained by the nutritional needs analysis module; meal compliance assessment module collects meal frequency, meal duration, average calorie intake, and calculated meal compliance index; meal adaptability assessment module assesses the meal adaptability index; and the patient's dietary behavior pattern type is identified by the dietary behavior pattern recognition module.

[0039] Compared with the prior art, the beneficial effects of the nutritional meal management system for stroke patients with dysphagia of the present invention are as follows:

[0040] 1. This invention integrates multiple functional modules to achieve comprehensive monitoring and intelligent management of the dietary status of stroke patients with dysphagia. The dietary recording module combines weight sensors and image recognition technology to improve the accuracy and convenience of dietary data collection; the symptom monitoring module's correlation analysis with dietary records uses mathematical models to quantify the strength of the association between food and symptoms, providing a scientific basis for patients to identify potential allergenic foods; the nutritional needs analysis module ensures the accuracy and timeliness of patients' nutritional intake through real-time monitoring and dynamic adjustment; the meal compliance and adaptation assessment module integrates nutritional intake index, symptom severity index, patient compliance index, and food intolerance index through a quantitative model, enabling evaluation of meal plans based on multi-dimensional patient data. This provides medical staff with objective and quantitative assessment indicators, facilitating timely adjustments to meal plans.

[0041] 2. This invention, based on a Long Short-Term Memory (LSTM) network algorithm, can identify and predict patients' long-term dietary behavior patterns, thereby optimizing future meal planning. This design not only provides the system with data analysis capabilities but also dynamically adjusts meal planning based on patients' past dietary behaviors to better match their health needs. Through this module, the system can achieve real-time adaptive adjustments, improving meal planning compliance and effectiveness, making patient dietary management more scientific and intelligent. Attached Figure Description

[0042] Figure 1 This is the interface of the medical management system in Reference 2.

[0043] Figure 2 This is a diagram of the healthy eating guidelines from Reference 2.

[0044] Figure 3 This is a schematic diagram of the structure of a nutritional meal management system for stroke patients with dysphagia according to the present invention.

[0045] In the image, HOME: Homepage; PATIENTS: Patients; ADMINISTRATIVE: Management; Add patient: Add patient; ID Card information: ID card information; Surname and name: Surname and first name; PIN: Personal Identification Number; Locality: Location; Address: Address; Phone: Telephone; Email: Email; Gender: Gender; Age: Age; Height: Height; Weight: Weight; Birthday: Birthday; Add patient: Add patient; You are logged in as admin: You are logged in as an administrator; Logout: Log out; Reduce your consumption of red meat, whole milk, butter, and eggs: Reduce your intake of red meat, whole milk, butter, and eggs, as these are major sources of saturated fat; Cook with olive oil instead of regular vegetable oil: Cook with olive oil instead of regular vegetable oil; rapeseed oil is also a good choice, especially suitable for baking; Check the ingredient list on food. Labels: Check the ingredient list on food labels and avoid foods containing hydrogenated or partially hydrogenated oils, which are commonly found in bar margarine, shortening, salad dressings, and other packaged foods; Trim the fat off meat: Trim the fat off meat and remove fat before eating, avoiding chicken skin; Choose nonfat dairy products: Choose nonfat dairy products and eggs fortified with Omega-3 fatty acids; Add nuts and seeds: Add nuts and seeds to cereals, salads, soups, or other dishes. Good choices include walnuts, almonds, pumpkin seeds, hazelnuts, pecans, and sesame seeds; Use flaxseed oil: Use flaxseed oil, which can be added to smoothies, salad dressings, or mixed into snacks such as applesauce, but do not heat flaxseed oil for cooking, as heating will cause it to lose its protective properties; Limit fast food, fried foods, and packaged foods: Limit your intake of fast food, fried foods, and packaged foods, which often contain high levels of trans fats, including potato chips, cookies, crackers, French fries, and donuts; Eat fish once or twice a day. week: Eat fish once or twice a week. Good choices include wild salmon, sardines, herring, and black cod; Nr.: Number. Detailed Implementation

[0046] The technical solutions of this embodiment will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0047] Example 1

[0048] A nutritional meal management system for stroke patients with dysphagia includes a patient information collection module, a diet record module, a symptom monitoring module, a nutritional needs analysis module, a meal compliance assessment module, a meal adaptation assessment module, a data storage module, and a dietary behavior pattern recognition module.

[0049] In this embodiment of the invention, the patient information collection module is connected to the data storage module and the nutritional requirements analysis module; the diet record module is connected to the data storage module, the symptom monitoring module, and the meal preparation adaptability assessment module; the symptom monitoring module is connected to the data storage module and the meal preparation adaptability assessment module; the nutritional requirements analysis module is connected to the data storage module and the meal preparation adaptability assessment module; the meal preparation compliance assessment module is connected to the data storage module and the meal preparation adaptability assessment module; the meal preparation adaptability assessment module is connected to the data storage module; and the dietary behavior pattern recognition module is connected to the data storage module, the meal preparation adaptability assessment module, the nutritional requirements analysis module, the meal preparation compliance assessment module, and the diet record module.

[0050] The patient information collection module is used to collect basic patient information, including age, gender, height, weight, disease stage, comorbidities, and allergy history.

[0051] The diet recording module includes a diet data acquisition unit, a symptom association recording unit, and a food intolerance index calculation unit. The diet data acquisition unit measures food weight using a weight sensor and uses image recognition technology to capture photos of the meal preparation, identify and record food types. It comprehensively utilizes weight data and image recognition results to record the types and weights of food consumed by the patient. The symptom association recording unit records the patient's daily symptoms, including symptom type, occurrence time, duration, and severity level. The food intolerance index calculation unit uses the data provided by the diet data acquisition unit and the symptom association recording unit to assess the patient's food intolerance index.

[0052] The diet recording module in this embodiment of the invention includes the following steps:

[0053] Step A1 involves analyzing the relationship between food intake weight and symptom presentation within a specific time window prior to symptom onset to quantify the association strength between food and symptoms. For each food and symptom, a food-symptom association matrix is ​​established, and its association index is calculated. The formula for calculating the association index between food and symptoms is as follows:

[0054]

[0055] In the formula, R fs Y represents the correlation index between food f and symptom s. s (τ) represents the severity rating of symptom s at time τ, where τ is the time point in time when the symptoms were recorded, and X f (τ-δ) represents the weight of food f ingested δ time before the onset of symptoms;

[0056] Step A2: Based on the food-symptom association matrix, construct a food intolerance index assessment model to evaluate the patient's degree of intolerance to the prepared food. The formula for the food intolerance index assessment model is as follows:

[0057]

[0058] In the formula, FII is the food intolerance index, F is the number of food types in the meal plan, S is the number of symptoms considered, and W is the number of symptoms considered. f V represents the weight of food f in the overall meal preparation, determined based on the food weight. s R represents the weight of symptom s. fs The correlation index between food f and symptom s.

[0059] The symptom monitoring module includes a symptom detection unit and a symptom feature extraction unit. The symptom monitoring unit is used to track and record changes in the patient's symptoms, including dysphagia, choking, choking while drinking, and weight changes, and transmits these symptoms to the symptom feature extraction unit. The symptom feature extraction unit is used to extract features from the data transmitted by the symptom monitoring unit to obtain the dysphagia index, choking frequency, choking frequency while drinking, and weight change percentage, and to construct a symptom severity assessment model to obtain a symptom severity index.

[0060] The formula for the symptom severity assessment model in this embodiment of the invention is as follows:

[0061]

[0062] In the formula, SSI is the symptom severity index, T is the monitoring period, AP(t) is the dysphagia index at time t, DF(t) is the choking frequency at time t, BL(t) is the choking frequency when drinking water at time t, and WC(t) is the percentage change in weight at time t.

[0063] The nutritional needs analysis module includes a nutritional needs customization unit, a nutritional intake monitoring unit, and a nutritional needs analysis unit. The nutritional needs customization unit records the patient's daily recommended nutrient intake. The nutritional intake monitoring unit monitors the patient's actual nutrient intake at each meal using a food nutrient composition monitoring and analysis instrument. The nutrients include calories, protein, fat, carbohydrates, vitamins, and minerals. The nutritional needs analysis unit assesses the patient's nutritional intake index by comparing the patient's actual nutrient intake with their recommended nutrient intake and constructing a nutritional intake index evaluation model.

[0064] The nutritional requirements analysis module in this embodiment of the invention includes the following steps:

[0065] Step B1: Obtain the patient's recommended daily nutrient intake from the nutrition needs customization unit;

[0066] Step B2: Obtain the patient's actual daily nutritional intake from the nutritional intake monitoring unit;

[0067] Step B3: Construct a nutritional intake index assessment model to obtain the patient's nutritional intake index. The formula for the nutritional intake index assessment model is as follows:

[0068]

[0069] In the formula, NII is the patient's nutritional intake index, n is the quantity of nutrients, and w i Let be the weight of the i-th nutrient component, where min() is the min function, A i T represents the actual daily nutrient intake. i This refers to the recommended daily intake of nutrients.

[0070] The meal adherence assessment module includes a diet execution monitoring unit and a adherence index assessment unit. The diet execution monitoring unit records the number of meals the patient actually eats each day, as well as the start and end times of each meal. It also measures the actual weight of food consumed by the patient at each meal using a weight sensor and, combined with a food nutrition database, obtains the average daily calorie intake. The adherence index assessment unit compares the number of meals the patient actually eats with the planned number of meals in the meal plan to assess adherence to meal frequency; compares the average calorie intake of the patient with the planned average calorie intake to analyze the implementation of calorie control; compares the average meal duration recorded by the patient with the planned average meal duration; and integrates these parameters into a quantitative adherence index to construct a patient adherence index assessment model and obtain the patient's meal adherence index.

[0071] The formula for the compliance index assessment model in this embodiment of the invention is as follows:

[0072]

[0073] In the formula, PCI is the patient's meal compliance index, ME is the number of meals actually performed, MP is the number of planned meals, AE is the average actual calorie intake, AP is the average planned calorie intake, TE is the average actual recorded meal duration, and TP is the average planned meal duration.

[0074] The meal preparation adaptability assessment module is used to construct a meal preparation adaptability assessment model based on the patient's nutritional intake index, symptom severity index, patient compliance index, and food intolerance index transmitted from the dietary record module, symptom monitoring module, nutritional requirements analysis module, and meal preparation compliance assessment module. This model assesses the adaptability of stroke patients with dysphagia to the current meal preparation plan. The formula for the meal preparation adaptability assessment model is as follows:

[0075]

[0076] In the formula, ADI is the meal fit index, NII is the nutritional intake index, SSI is the symptom severity index, PCI is the patient compliance index, and FII is the food intolerance index.

[0077] The dietary behavior pattern recognition module is used to identify patients' long-term dietary behavior patterns using a long short-term memory network algorithm.

[0078] The dietary behavior pattern recognition module in this embodiment of the invention includes the following steps:

[0079] Step S1: Construct a dataset for identifying dietary behavior patterns. The dataset includes a meal fit index, a nutrient intake index, a patient compliance index, and a food intolerance index. The meal fit index is calculated by the meal fit assessment module; the nutrient intake index is calculated by the nutrient requirements analysis module; the patient compliance index is calculated by the meal compliance assessment module; and the food intolerance index is calculated by the dietary record module. The dataset covers the patient's daily data over the past 6 months.

[0080] Step S2 involves preprocessing the dataset, including organizing the daily data into a time series format, standardizing each index, constructing a sliding window with a window size of 30 days and a step size of 1 day, and dividing the preprocessed dataset into training, validation, and test sets in an 8:1:1 ratio.

[0081] Step S3: Construct a long short-term memory network model, which includes an input layer, two long short-term memory network layers, two discard layers, and a fully connected layer; the input layer receives meal fit index, nutrient intake index, patient compliance index, and food intolerance index.

[0082] Step S4: Train the Long Short-Term Memory (LSTM) network model using the training set data, including using the Adam optimizer, setting the learning rate to 0.001, selecting mean squared error as the loss function, setting the batch size to 32, setting the number of training epochs to 100, and using an early stopping strategy with a tolerance of 10 epochs; monitor the model performance on the validation set.

[0083] Step S5: Evaluate the model performance on the test set, calculate the root mean square error and mean absolute error for each index, and assess the model's prediction accuracy; use the trained model to predict the index for the next 30 days, analyze the prediction results, and output the identified patient dietary behavior pattern types, including stable, fluctuating, improving, worsening, and mixed types.

[0084] Below is a Python code example that uses a long short-term memory network model to identify patient eating behavior patterns. Please note that this example is only a starting point and may need to be adjusted according to the actual situation and device interface in real applications.

[0085]

[0086]

[0087]

[0088]

[0089] This code is for illustrative purposes only. In actual applications, it needs to be modified and adjusted according to specific circumstances.

[0090] In this embodiment of the invention, the first long short-term memory network layer contains 64 units and the activation function is tanh; the second long short-term memory network layer contains 32 units and the activation function is tanh; the discard rate of both discard layers is 0.2; and the fully connected layer contains 4 units and the activation function is a linear function.

[0091] In this embodiment of the invention, the stable type is defined as all indices changing by less than 10%; the volatile type is defined as any index changing by more than 30%; the improving type is defined as ADI and NII showing an upward trend, and FII showing a downward trend; the deteriorating type is defined as ADI and NII showing a downward trend, and FII showing an upward trend; and the mixed type is defined as a pattern that does not conform to any of the above types.

[0092] The data storage module stores data from various modules, including basic information collected by the patient information collection module; food types, weights, and food intolerance indices recorded by the diet record module; dysphagia index, choking frequency, choking frequency when drinking water, percentage weight change, and symptom severity index obtained by the symptom monitoring module; nutrient intake and nutrient intake index obtained by the nutritional needs analysis module; meal compliance assessment module collecting meal frequency, meal duration, average calorie intake, and calculated meal compliance index; meal adaptability assessment module evaluating the meal adaptability index; and patient dietary behavior pattern types identified by the dietary behavior pattern recognition module.

[0093] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0094] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A nutritional meal management system for stroke patients with dysphagia, comprising a patient information collection module, a dietary record module, a symptom monitoring module, a nutritional needs analysis module, a meal compliance assessment module, a meal adaptation assessment module, a data storage module, and a dietary behavior pattern recognition module, characterized in that, The patient information collection module is connected to the data storage module and the nutritional requirements analysis module; the diet record module is connected to the data storage module, the symptom monitoring module, and the meal preparation adaptability assessment module; the symptom monitoring module is connected to the data storage module and the meal preparation adaptability assessment module; the nutritional requirements analysis module is connected to the data storage module and the meal preparation adaptability assessment module; the meal preparation compliance assessment module is connected to the data storage module and the meal preparation adaptability assessment module; the meal preparation adaptability assessment module is connected to the data storage module; the dietary behavior pattern recognition module is connected to the data storage module, the meal preparation adaptability assessment module, the nutritional requirements analysis module, the meal preparation compliance assessment module, and the diet record module; the meal preparation adaptability assessment module is used to construct a meal preparation adaptability assessment model based on the patient nutritional intake index, symptom severity index, patient compliance index, and food intolerance index transmitted by the diet record module, the symptom monitoring module, the nutritional requirements analysis module, and the meal preparation compliance assessment module, to assess the adaptability index of stroke patients with dysphagia to the current meal plan, wherein the formula of the meal preparation adaptability assessment model is: In the formula, ADI is the meal fit index, NII is the nutritional intake index, SSI is the symptom severity index, PCI is the patient compliance index, and FII is the food intolerance index.

2. A nutritional meal management system for stroke patients with dysphagia according to claim 1, characterized in that, The diet recording module includes a diet data acquisition unit, a symptom association recording unit, and a food intolerance index calculation unit. The diet data acquisition unit measures food weight using a weight sensor and uses image recognition technology to capture photos of the meal preparation, identify, and record food types. The diet data acquisition unit comprehensively utilizes weight data and image recognition results to record the types and weights of food consumed by the patient. The symptom association recording unit records the patient's daily symptoms, including symptom type, occurrence time, duration, and severity level. The food intolerance index calculation unit uses the data provided by the diet data acquisition unit and the symptom association recording unit to assess the patient's food intolerance index, including the following steps: Step A1 involves analyzing the relationship between food intake weight and symptom presentation within a specific time window prior to symptom onset to quantify the association strength between food and symptoms. For each food and symptom, a food-symptom association matrix is ​​established, and its association index is calculated. The formula for calculating the association index between food and symptoms is as follows: In the formula, R fs Y represents the correlation index between food f and symptom s. s (τ) represents the severity rating of symptom s at time τ, where τ is the time point in time when the symptoms were recorded, and X f (τ-δ) represents the weight of food f ingested δ time before the onset of symptoms; Step A2: Based on the food-symptom association matrix, construct a food intolerance index assessment model to evaluate the patient's degree of intolerance to the prepared food. The formula for the food intolerance index assessment model is as follows: In the formula, FII is the food intolerance index, F is the number of food types in the meal plan, S is the number of symptoms considered, and W is the number of symptoms considered. f V represents the weight of food f in the overall meal preparation, determined based on the food weight. s R represents the weight of symptom s. fs The correlation index between food f and symptom s.

3. A nutritional meal management system for stroke patients with dysphagia according to claim 1, characterized in that, The symptom monitoring module includes a symptom detection unit and a symptom feature extraction unit. The symptom monitoring unit is used to track and record changes in the patient's symptoms, including symptoms such as difficulty swallowing, choking, choking while drinking, and weight changes, and transmits them to the symptom feature extraction unit. The symptom feature extraction unit is used to extract features from the data transmitted by the symptom monitoring unit, obtaining the dysphagia index, choking frequency, choking frequency when drinking water, and percentage change in weight. It also constructs a symptom severity assessment model to obtain the symptom severity index. The formula for the symptom severity assessment model is as follows: In the formula, SSI is the symptom severity index, T is the monitoring period, AP(t) is the dysphagia index at time t, DF(t) is the choking frequency at time t, BL(t) is the choking frequency when drinking water at time t, and WC(t) is the percentage change in weight at time t.

4. A nutritional meal management system for stroke patients with dysphagia according to claim 1, characterized in that, The nutrition requirements analysis module includes a nutrition requirements customization unit, a nutrition intake monitoring unit, and a nutrition requirements analysis unit. The nutritional requirement customization unit is used to record the recommended intake of nutrients for each patient's purpose. The nutrient intake monitoring unit is used to monitor the patient's actual nutrient intake at each meal using a food nutrient composition monitoring and analysis instrument; the nutrients include calories, protein, fat, carbohydrates, vitamins, and minerals; the nutrient requirement analysis unit is used to construct a nutrient intake index assessment model by comparing the patient's actual nutrient intake with their recommended nutrient intake, and to assess the patient's nutrient intake index, including the following steps: Step B1: Obtain the patient's recommended daily nutrient intake from the nutrition needs customization unit; Step B2: Obtain the patient's actual daily nutritional intake from the nutritional intake monitoring unit; Step B3: Construct a nutritional intake index assessment model to obtain the patient's nutritional intake index. The formula for the nutritional intake index assessment model is as follows: In the formula, NII is the patient's nutritional intake index, n is the quantity of nutrients, and w i Let be the weight of the i-th nutrient component, where min() is the min function, A i T represents the actual daily nutrient intake. i This refers to the recommended daily intake of nutrients.

5. A nutritional meal management system for stroke patients with dysphagia according to claim 1, characterized in that, The meal adherence assessment module includes a diet execution monitoring unit and a adherence index assessment unit. The diet execution monitoring unit records the number of meals the patient actually eats each day and the start and end times of each meal. It also measures the actual weight of food consumed by the patient at each meal using a weight sensor and, in conjunction with a food nutrition database, obtains the average daily food calorie intake. The adherence index assessment unit compares the number of meals the patient actually eats with the planned number of meals in the meal plan to assess the degree of adherence to the meal frequency. Compare the patients' actual average calorie intake with their planned average calorie intake to analyze the implementation of calorie control; compare the patients' actual recorded average meal duration with their planned average meal duration. The various parameters were integrated into a quantitative adherence index to construct a patient adherence index assessment model, obtaining the patient's meal adherence index. The formula for the adherence index assessment model is as follows: In the formula, PCI is the patient's meal compliance index, ME is the number of meals actually performed, MP is the number of planned meals, AE is the average actual calorie intake, AP is the average planned calorie intake, TE is the average actual recorded meal duration, and TP is the average planned meal duration.

6. A nutritional meal management system for stroke patients with dysphagia according to claim 1, characterized in that, The dietary behavior pattern recognition module is used to identify patients' long-term dietary behavior patterns using a long short-term memory network algorithm, including the following steps: Step S1: Construct a dataset for identifying dietary behavior patterns. The dataset includes a meal fit index, a nutrient intake index, a patient compliance index, and a food intolerance index. The meal fit index is calculated by the meal fit assessment module; the nutrient intake index is calculated by the nutrient requirements analysis module; the patient compliance index is calculated by the meal compliance assessment module; and the food intolerance index is calculated by the dietary record module. The dataset covers the patient's daily data over the past 6 months. Step S2 involves preprocessing the dataset, including organizing the daily data into a time series format, standardizing each index, constructing a sliding window with a window size of 30 days and a step size of 1 day, and dividing the preprocessed dataset into training, validation, and test sets in a ratio of 8:1:

1. Step S3: Construct a long short-term memory network model, which includes an input layer, a long short-term memory network layer, a discard layer, and a fully connected layer; the input layer receives meal fit index, nutrient intake index, patient compliance index, and food intolerance index. Step S4: Train the Long Short-Term Memory network model using the training set data, and monitor the model performance on the validation set using the Adam optimizer and early stopping strategy. Step S5: Evaluate the model performance on the test set, calculate the root mean square error and mean absolute error for each index, and assess the model's prediction accuracy; use the trained model to predict the index for the next 30 days, analyze the prediction results, and output the identified patient dietary behavior pattern type.

7. A nutritional meal management system for stroke patients with dysphagia according to claim 1, characterized in that, The patient information collection module is used to collect basic patient information, including age, gender, height, weight, disease stage, comorbidities, and allergy history; the data storage module is used to store data from various modules, including basic information collected by the patient information collection module, food types, weights, and food intolerance index recorded by the diet record module; and dysphagia index, choking frequency, choking frequency when drinking water, percentage change in weight, and symptom severity index obtained by the symptom monitoring module. The nutritional needs analysis module obtains the intake of nutrients and the nutritional intake index; the meal compliance assessment module collects the number of meals, meal duration, average calorie intake, and calculates the meal compliance index; the meal fit assessment module assesses the meal fit index; and the dietary behavior pattern recognition module identifies the types of patient dietary behavior patterns.

Images