Nutrition evaluation system based on multi-source data

Abstract

The invention relates to the technical field of nutrition data analysis, in particular to a nutrition evaluation system based on multi-source data, which comprises a parameter extraction module, a parameter adjustment module, a parameter screening module, a conversion calculation module and an evaluation generation module. According to the method, the individualized gastric emptying rate and amplitude are accurately quantified by acquiring the stomach image in real time and analyzing the volume change of the stomach image at continuous time points, and individualized weight correction is carried out on conventional plasma digestive enzyme activity data by taking the dynamic digestive function parameter as a regulatory factor; then, in combination with instant heart rate and stride frequency data collected by the wearable device, the instant metabolic state of the user is judged, the corrected enzyme activity parameters are screened and matched according to the state, and a personalized nutrition absorption capacity index highly related to the current physiological demand is deduced; and finally, by comparing the dynamic differences of the nutrient conversion amounts at different time points, comprehensive evaluation of the dynamic stability of the nutrient absorption efficiency and the component utilization adaptability is realized.

Description

Technical Field

[0001] This invention relates to the field of nutritional data analysis technology, and in particular to a nutritional assessment system based on multi-source data. Background Technology

[0002] The field of nutritional data analysis technology involves technical means to assess, monitor, and intervene in human nutritional status. It mainly involves analyzing the nutritional intake, nutritional needs, and nutritional status of individuals or groups to ensure a scientific and reasonable dietary structure.

[0003] Among them, a nutrition assessment system based on multi-source data refers to a system that integrates data from different sources, such as patients' enteral nutrition intake data, health check-up data, laboratory test data, and clinical treatment information, to comprehensively analyze and assess an individual's nutritional status.

[0004] Existing technologies merely integrate and list heterogeneous data from multiple sources, failing to reveal the intrinsic connections between various physiological indicators. For example, they cannot effectively correlate patients' laboratory test data, such as digestive enzyme activity, with their actual gastrointestinal motility characteristics. They also ignore the dynamic impact of changes in an individual's real-time physiological state, such as metabolic levels, on nutrient absorption efficiency at different times. This assessment method is essentially static and fragmented, and its results often lack timeliness and contextual relevance. For instance, for a user who is in a quiet office state in the afternoon and engages in light activity at night, existing technologies may fail to capture this metabolic difference, leading to a vague conclusion that the absorption capacity of nutrients is the same in both instances. This could result in a mismatch between nutrient supply and actual physiological needs, affecting the accuracy and effectiveness of nutritional support. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by proposing a nutrition assessment system based on multi-source data.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a nutrition assessment system based on multi-source data includes: The parameter extraction module acquires ultrasound images of the user's stomach at a specified time, analyzes the volume of gastric contents at each time, calculates the rate of descent of gastric contents over continuous time, calculates the total gastric emptying amplitude by calculating the volume difference between the initial and final gastric contents, and constructs a set of gastric emptying rate parameters. The parameter adjustment module acquires pancreatic digestive enzyme activity data from the user's fasting blood sample, and adjusts the weights of the pancreatic digestive enzyme activity data in conjunction with the gastric emptying rate parameter set to generate an enzyme activity adjustment parameter set. The parameter filtering module obtains the user's daily step frequency data and average heart rate, filters nutrient absorption parameters that match the metabolic state from the enzyme activity regulation parameter group, and outputs the absorption adaptation parameter group. The conversion calculation module obtains the protein, carbohydrate and lipid content per 100 ml of the enteral nutrition preparation ingested by the user, and calculates the conversion amount of the three types of parameters through the absorption adaptability parameter group to obtain nutrient conversion structure data. The assessment generation module acquires the nutrient conversion structure data after multiple enteral nutrition intakes by the user, calculates the difference in conversion amount in chronological order, judges the status of enteral nutrient absorption capacity, and obtains the enteral nutrition assessment result.

[0007] As a further aspect of the present invention, the gastric emptying rate parameter set includes the rate of descent of gastric contents and the total emptying amplitude of gastric contents; the enzyme activity regulation parameter set includes adjusted enzyme activity values, digestive pathway correspondences, and component priority order; the absorption adaptation parameter set includes intraday metabolic status labels, intake behavior characteristics, and enzyme activity selection range; the nutrient conversion structure data includes protein conversion amount, carbohydrate conversion amount, and lipid conversion amount; and the enteral nutrition assessment results include protein absorption trend status, carbohydrate absorption trend status, and lipid absorption trend status.

[0008] As a further aspect of the present invention, the parameter extraction module includes: The image acquisition submodule acquires ultrasound images of the stomach at multiple time points when the user takes oral glucose solution, extracts the anteroposterior diameter boundary of the sagittal section region of the stomach cavity in the image, and locates and marks the image through image frame sequence and time label to generate anteroposterior diameter localization data. The volume analysis submodule, based on the anteroposterior diameter positioning data, calls the anteroposterior diameter change value in the image at each time point, combines the image shooting angle parameter and user body position information, and uses a geometric fitting method to establish the volume estimation relationship of the contents, generating a sequence of gastric contents volume values ​​at each time point; The emptying calculation submodule calls the sequence of gastric contents volume values ​​at each time point, calculates the rate of gastric contents descent over a continuous period, and calculates the volume difference between the initial volume and the final volume as the total gastric contents emptying amplitude, generating a gastric emptying rate parameter set.

[0009] As a further aspect of the present invention, the user takes oral glucose solution at multiple time points, namely 20 minutes, 40 minutes, and 60 minutes.

[0010] As a further aspect of the present invention, the parameter adjustment module includes: The enzyme activity acquisition submodule acquires venous blood samples taken by the user in a fasting state, uses colorimetric method to measure the activity data of each type of pancreatic digestive enzyme, timestamps and identifies the measured values ​​with project numbers, and establishes a pancreatic digestive enzyme activity dataset. The parameter normalization submodule extracts the gastric contents descent rate and the total gastric contents emptying amplitude from the gastric emptying rate parameter group, performs normalization on the two parameters, and outputs normalized emptying parameter data. The weight adjustment submodule adjusts the weights of each type of pancreatic digestive enzyme activity data based on the pancreatic digestive enzyme activity dataset and the normalized emptying parameter data, respectively, to generate an enzyme activity regulation parameter set.

[0011] As a further aspect of the present invention, the pancreatic digestive enzyme activity data includes the activity values ​​of pancreatic lipase, pancreatic amylase and trypsin.

[0012] As a further aspect of the present invention, the parameter filtering module includes: The physiological parameter acquisition submodule acquires continuous cadence data recorded by the user's wearable device on the same day, calculates the range of cadence variation throughout the day, and combines the average heart rate values ​​during rest and activity periods. It then uses a time axis positioning method to synchronize and organize cadence and heart rate data, generating a combined daily exercise and heart rate data set. The metabolic state recognition submodule calls the combined daily exercise and heart rate data set, extracts the step frequency change amplitude and average heart rate value, calculates the metabolic fluctuation value based on the linkage relationship between the two values, and determines the current metabolic state based on the set daily metabolic state recognition interval, and assigns metabolic state classification labels. The parameter screening and matching submodule, based on the enzyme activity regulation parameter group and the metabolic state classification label, filters the nutrient absorption parameters corresponding to the current metabolic state from the enzyme activity parameters, including pancreatic lipase absorption parameters, pancreatic amylase absorption parameters and trypsin absorption parameters, and outputs the absorption adaptation parameter group.

[0013] As a further aspect of the present invention, the conversion calculation module includes: The component parameter extraction submodule obtains the label information of the enteral nutrition preparation ingested by the user, extracts the labeled content values ​​of protein, carbohydrates and lipids per 100 ml, and generates a standard component content parameter set. The data absorption call submodule matches the screened nutrient absorption parameters in the absorption adaptability parameter group, establishes an association table with the corresponding three types of nutrients, and generates an absorption and conversion correspondence group. The transformation structure generation submodule, based on the standard component content parameter group and the absorption and transformation correspondence group, transforms proteins, carbohydrates and lipids respectively, including protein transformation amount, carbohydrate transformation amount and lipid transformation amount, and outputs nutrient transformation structure data.

[0014] As a further aspect of the present invention, the evaluation generation module includes: The intake record integration submodule obtains the time points of each user's enteral nutrient intake, arranges each time point in sequence, calls the nutrient conversion structure data corresponding to each time point, and generates an intake conversion time series data group. The conversion difference calculation submodule calls the intake conversion time series data group, extracts the conversion amount of protein, carbohydrates and lipids between any two adjacent intakes, calculates the conversion amount difference respectively and performs component classification and combination to generate nutrient component conversion difference combination data. The absorption status determination submodule determines the changing trends of protein, carbohydrate and lipid conversion based on the nutrient conversion difference combination data, compares them with the preset absorption stability range, and outputs the enteral nutrition assessment results for a comprehensive assessment of enteral nutrition absorption efficiency, dynamic stability and component utilization.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, by acquiring real-time images of the stomach and analyzing changes in its contents at continuous time points, the individualized gastric emptying rate and amplitude are precisely quantified. This dynamic digestive function parameter is used as a regulatory factor to individually weight and correct conventional plasma digestive enzyme activity data. Combined with real-time heart rate and cadence data collected by wearable devices, the user's real-time metabolic state is determined. Based on this state, the corrected enzyme activity parameters are screened and matched to derive a personalized nutrient absorption capacity index that is highly correlated with current physiological needs. Finally, by comparing the dynamic differences in nutrient conversion at different time points, a comprehensive assessment of the dynamic stability of nutrient absorption efficiency and the adaptability of component utilization is achieved, providing an objective basis for developing more timely and individually adaptable nutritional intervention programs. Attached Figure Description

[0016] Figure 1 This is a system flowchart of the present invention; Figure 2 This is a flowchart of the parameter extraction module of the present invention; Figure 3 This is a flowchart of the parameter adjustment module of the present invention; Figure 4 This is a flowchart of the parameter filtering module of the present invention; Figure 5 This is a flowchart of the conversion calculation module of the present invention; Figure 6 This is a flowchart of the evaluation and generation module of the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0018] Please see Figure 1 A nutrition assessment system based on multi-source data includes: The parameter extraction module acquires ultrasound images of the user's stomach at a specified time, analyzes the volume of gastric contents at each time, calculates the rate of descent of gastric contents over continuous time, calculates the total gastric emptying amplitude by calculating the volume difference between the initial and final gastric contents, and constructs a set of gastric emptying rate parameters. The parameter adjustment module acquires pancreatic digestive enzyme activity data from the user's fasting blood sample, and adjusts the weights of the pancreatic digestive enzyme activity data in conjunction with the gastric emptying rate parameter set to generate an enzyme activity adjustment parameter set. The parameter filtering module obtains the user's daily step frequency data and average heart rate, filters nutrient absorption parameters that match the metabolic state from the enzyme activity regulation parameter group, and outputs the absorption adaptation parameter group. The conversion calculation module obtains the protein, carbohydrate and lipid content per 100 ml of the enteral nutrition preparation ingested by the user, and calculates the conversion amount of the three types of parameters by deriving the absorption adaptability parameter set to obtain nutrient conversion structure data. The assessment generation module acquires nutrient conversion structure data after multiple enteral nutrition intakes by the user, calculates the difference in conversion amount in chronological order, judges the status of enteral nutrient absorption capacity, and obtains enteral nutrition assessment results. The gastric emptying rate parameter group includes the rate of descent of gastric contents and the total amount of gastric contents emptied. The enzyme activity regulation parameter group includes adjusted enzyme activity values, digestive pathway correspondences, and component priority order. The absorption adaptation parameter group includes intraday metabolic status labels, intake behavior characteristics, and enzyme activity selection range. The nutrient conversion structure data includes protein conversion, carbohydrate conversion, and lipid conversion. The enteral nutrition assessment results include protein absorption trend status, carbohydrate absorption trend status, and lipid absorption trend status.

[0019] Please see Figure 2 The parameter extraction module includes: The image acquisition submodule acquires gastric ultrasound images of the user at multiple time points, including 20, 40, and 60 minutes after oral glucose administration. It extracts the anteroposterior diameter boundary of the sagittal section of the gastric cavity in the images, and uses image frame sequences and time labels to locate and label the images, generating anteroposterior diameter localization data. After the user orally ingested 250 ml of a standard liquid containing 75 grams of glucose, ultrasound images of the user's stomach were acquired at three preset time points: 20 minutes, 40 minutes, and 60 minutes. Specifically, at the 20-minute mark, the operator placed the ultrasound probe coated with coupling gel on the midline below the xiphoid process. By slightly adjusting the probe angle and pressure, a sagittal section image showing the maximum anteroposterior diameter of the gastric antrum was obtained on the screen. The playback function of the device was then activated, continuously recording a 5-second dynamic image sequence totaling 125 frames. The operator reviewed the dynamic image sequence frame by frame and selected the frame showing the largest distance between the anterior and posterior walls of the stomach (frame 88) for analysis, using the electronic calipers built into the ultrasound diagnostic instrument. The anteroposterior diameter was measured vertically from the inner edge of the anterior wall mucosa to the inner edge of the posterior wall mucosa, and the maximum vertical distance was precisely recorded. The anteroposterior diameter obtained in this measurement was 25.0 mm. At the same time, the current acquisition timestamp "September 5th, 10:20:10" was bound to the unique identifier of the image frame "Seq01-Frame88" as the image positioning and marking information. Following this rigorous method, the anteroposterior diameter was measured to be 18.0 mm at 40 minutes and 11.5 mm at 60 minutes. The set of measurement data from the three time points [{Time point: 20 minutes, anteroposterior diameter: 25.0 mm}, {Time point: 40 minutes, anteroposterior diameter: 18.0 mm}, {Time point: 60 minutes, anteroposterior diameter: 11.5 mm}] was used to generate the anteroposterior diameter positioning data.

[0020] The volume analysis submodule, based on the anteroposterior diameter positioning data, calls the anteroposterior diameter change values ​​in the images at each time point, combines the image shooting angle parameters and user body position information, and uses a geometric fitting method to establish the volume estimation relationship of the contents, generating a sequence of gastric contents volume values ​​at each time point; Using anteroposterior diameter positioning data at [{time point: 20 minutes, anteroposterior diameter: 25.0 mm}, {time point: 40 minutes, anteroposterior diameter: 18.0 mm}, and {time point: 60 minutes, anteroposterior diameter: 11.5 mm}], combined with the user's guided and maintained standard supine position throughout the measurement process, and the fixed imaging angle parameter of maintaining 30 degrees ± 2 degrees between the ultrasound probe and the user's abdominal midline, a preset and validated geometric fitting relationship was used to estimate the instantaneous volume of gastric contents. This geometric fitting relationship was established based on a double-blind study involving 500 healthy adult volunteers who underwent simultaneous gastric ultrasound and high-resolution magnetic resonance imaging (MRI) measurements. The study established a mathematical correlation model between the anteroposterior diameter of the gastric antrum measured by ultrasound and the volume of gastric contents accurately measured by MRI by performing nonlinear regression analysis on the two sets of imaging data. Specifically, the estimation method approximates the sagittal section of the gastric cavity as an ellipse, with the corresponding transverse diameter set to 1.5 times the anteroposterior diameter based on the model's statistical results. The ellipse area is then multiplied by an effective length corrected for the body surface area (1.83 square meters) calculated based on the user's height (1.75 meters) and weight (70 kilograms). In this example, the effective length is 52 millimeters. Therefore, at the 20-minute mark, the anteroposterior diameter is 25.0 millimeters, the corresponding transverse diameter is 37.5 millimeters, and the estimated gastric contents volume is... The calculated result is 38289 cubic millimeters, or 38.3 milliliters. Using the same method, the volume at the 40th minute is calculated to be 19.9 milliliters, and the volume at the 60th minute is 8.2 milliliters. The three calculation results are arranged in chronological order to generate a sequence of gastric contents volume values ​​at each time point.

[0021] The emptying calculation submodule calls the sequence of gastric contents volume values ​​at each time point, calculates the rate of gastric contents descent over a continuous period, calculates the volume difference between the initial volume and the final volume as the total gastric contents emptying amplitude, and generates a set of gastric emptying speed parameters. Using the gastric contents volume sequence [{20 minutes: 38.3 ml}, {40 minutes: 19.9 ml}, {60 minutes: 8.2 ml}], firstly, the rate of decrease in gastric contents during the first time period (20 minutes to 40 minutes) was calculated by subtracting the volume values ​​at the two time points and dividing by the time interval, i.e., (38.3 ml - 19.9 ml) / (40 minutes - 20 minutes), yielding a rate of 0.92 ml / min for this period. Then, the rate of decrease during the second time period (40 minutes to 60 minutes) was calculated by (19.9 ml - 8.2 ml) / (60 minutes - 40 minutes), yielding a rate of 0.585 ml / min for this period. This process was used to obtain an energy... The table shows the rate of gastric contents descent over a continuous period throughout the entire monitoring cycle. The rates from the two time periods are arithmetically averaged, i.e., (0.92 + 0.585) / 2, resulting in a final average descent rate of 0.7525 ml / min. In addition, the emptying calculation submodule calculates the volume at the starting time point (20 minutes) of 38.3 ml and the volume at the last time point (60 minutes) of 8.2 ml, and calculates the absolute volume difference between the two, i.e., 38.3 ml - 8.2 ml, resulting in 30.1 ml. This value is taken as the total gastric contents emptying amplitude. Finally, the gastric contents descent rate of 0.7525 ml / min and the total gastric contents emptying amplitude of 30.1 ml are integrated to generate a gastric emptying rate parameter set.

[0022] Please see Figure 3 The parameter adjustment module includes: The enzyme activity acquisition submodule acquires venous blood samples taken by the user in a fasting state, uses colorimetric method to measure the activity data of each type of pancreatic digestive enzyme, timestamps and identifies the measured values ​​with project numbers, and establishes a pancreatic digestive enzyme activity dataset. At 7:30 AM on the day the user underwent the gastric emptying test, after confirming through questioning that the user had strictly fasted for more than 12 hours, a qualified medical professional, following aseptic procedures, drew 5 ml of venous blood from the user's median cubital vein. The blood sample was injected into a vacuum blood collection tube containing ethylenediaminetetraacetic acid (EDTA) as an anticoagulant and transported to the laboratory within 30 minutes. The sample was immediately centrifuged at 4000 rpm for 10 minutes at 4°C. The supernatant plasma was precisely aspirated and aliquoted into three independent cryovials. Using a fully automated biochemical analyzer, the activities of three pancreatic digestive enzymes were measured according to the reference methods recommended by the International Federation for Clinical Chemistry (IFCC). The pancreatic lipase was measured using 1,2-di(2,3 ... Using glycerol-3-valeric acid-(6-methylresorcinol) ester as a substrate, the reaction rate was monitored at a wavelength of 580 nm, and the measured activity value was 62 U / L. Pancreatic amylase was measured using ethylidene-4-nitrobenzene-G7 as a substrate, and the measured activity value was 95 U / L at a wavelength of 405 nm. Trypsin was measured using Nα-benzoyl-L-arginine-p-nitroaniline as a substrate, and the measured activity value was 53 U / L at a wavelength of 410 nm. The three measurements were time-stamped as "September 5th, 07:30" and assigned unique sample numbers "LIP250905", "AMY250905", and "TRY250905" respectively, thus establishing a dataset of pancreatic digestive enzyme activity.

[0023] The parameter normalization submodule extracts the rate of descent of gastric contents and the total emptying amplitude of gastric contents from the gastric emptying rate parameter group, performs normalization on the two parameters, and outputs normalized emptying parameter data. First, two parameters were extracted from the gastric emptying rate parameter set: the rate of descent of gastric contents (0.7525 ml / min) and the total emptying volume of gastric contents (30.1 ml). Then, minimum-maximum normalization was performed on these two parameters. This normalization aims to convert parameters of different dimensions to the same [0, 1] interval. The maximum and minimum value intervals for normalization were determined by statistically analyzing the results of a standardized gastric emptying test performed on a group of two thousand healthy adults with no history of gastrointestinal or metabolic diseases and a body mass index (BMI) between 18.5 and 24.0, after removing 5% of the extreme values. Based on established physiological normal ranges, this study defined the normal range for the rate of descent of gastric contents as 0.20 ml / min to 1.50 ml / min, and the normal range for the total emptying amplitude as 15.0 ml to 50.0 ml. The normalized value for the descent rate was (0.7525-0.20) / (1.50-0.20), which was calculated to be 0.425. The normalized value for the total emptying amplitude was (30.1-15.0) / (50.0-15.0), which was calculated to be 0.431. The two dimensionless values ​​were combined to output the normalized emptying parameter data: {Normalized descent rate: 0.425, Normalized total emptying amplitude: 0.431}.

[0024] The weight adjustment submodule uses the pancreatic digestive enzyme activity dataset and the normalized emptying parameter data as regulating factors to adjust the weights of each type of pancreatic digestive enzyme activity data, generating a set of enzyme activity regulation parameters. Using the pancreatic digestive enzyme activity dataset {pancreatic lipase: 62 U / L, pancreatic amylase: 95 U / L, trypsin: 53 U / L} and normalized emptying parameter data {normalized rate of decrease: 0.425, normalized total emptying magnitude: 0.431}, individualized weight adjustments were made to the activity values ​​of the three enzyme classes. The adjustment calculation process is described by the formula... The meaning and value determination process of each parameter in the formula are given below: The weighted parameter representing a specific enzyme E (i.e., pancreatic lipase, pancreatic amylase, or trypsin) after adjustment based on the user's individual gastric motility characteristics is the ultimate indicator for assessing potential digestive capacity. This represents the raw, measured activity of enzyme E in the user's plasma, reflecting the pancreas's basal secretory capacity, such as for pancreatic lipase. The value is 62 U / L. This is the normalized rate of decrease, representing the speed of gastric emptying, with a value of 0.425. This represents the normalized total emptying margin, which is 0.431, indicating the gastric emptying volume. and This is a dimensionless adjustment coefficient. The specific value of the adjustment coefficient is based on a large-scale prospective clinical study. This study used isotope labeling to accurately measure the absorption rates of lipids, carbohydrates, and proteins in different individuals, and performed multiple linear regression analysis with gastric emptying parameters. This quantified the impact of gastric emptying rate and amplitude on the digestion of various nutrients. The rationale is that lipid digestion is slow and inhibits gastric emptying; therefore, the total emptying volume (amplitude) better reflects digestive potential, and thus has a greater impact on the amplitude. The coefficient was set to a relatively large 0.3, while the rate was affected by The coefficient is 0.1; carbohydrates are digested quickly and are more directly affected by the emptying rate, therefore the factor affecting the rate is... The coefficient was set to a relatively large 0.2, which affected the magnitude of the influence. The coefficient is 0.1; protein digestion falls between these two extremes, and is relatively evenly affected by both rate and magnitude of change. and All coefficients are set to 0.15. Substituting the values ​​into the formula for complete calculation, the regulatory parameter for pancreatic lipase is: The regulatory parameters of pancreatic amylase are: The regulatory parameters of trypsin are The final set of enzyme activity regulation parameters was generated: {Adjusted lipase parameter: 72.65, adjusted amylase parameter: 107.17, adjusted protease parameter: 59.81}.

[0025] Please see Figure 4 The parameter filtering module includes: The physiological parameter acquisition submodule acquires continuous cadence data recorded by the user's wearable device on the same day, calculates the range of cadence variation throughout the day, and combines the average heart rate values ​​during rest and activity periods. It then uses a time axis positioning method to synchronize and organize cadence and heart rate data, generating a combined daily exercise and heart rate data set. A smart wristband-style wearable device, continuously worn by the user from waking up until the current moment, with a built-in three-axis accelerometer and photoplethysmography (PPG) sensor, continuously records and stores the user's cadence and heart rate data at 1-minute sampling intervals. For example, at 10:15 AM, when the user walks a short distance in the office, the device records a cadence of 110 steps / minute and a heart rate of 105 beats / minute. At 3:05 PM, when the user sits still in a meeting, the recorded cadence is 0 steps / minute and the heart rate is 72 beats / minute. By iterating through all the data points recorded for the day, the maximum cadence for the day is determined to be 140 steps / minute (occurring during a brisk walk at midday), and the minimum is 0 steps / minute. The device calculates the daily cadence range to be 140 steps per minute. Simultaneously, it records the user's average heart rate during the resting sleep period from 3:00 AM to 5:00 AM, with an average heart rate of 58 beats per minute. The device automatically synchronizes and organizes 300 heart rate data points and 5 cadence data points within a five-minute time window, calculating and recording the average cadence and average heart rate for each window. For example, the data points for the "10:15-10:20" time window are organized as {average cadence: 95 steps / minute, average heart rate: 102 beats / minute}. All the generated five-minute time window data for the entire day are integrated to generate a combined daily exercise and heart rate data set.

[0026] The metabolic state recognition submodule calls the combined daily exercise and heart rate data set, extracts the step frequency change amplitude and average heart rate value, calculates the metabolic fluctuation value based on the linkage between the two values, and determines the current metabolic state based on the set daily metabolic state recognition interval, and assigns metabolic state classification labels. By accessing the combined daily exercise and heart rate data set, at the precise moment before the user plans their next enteral nutrition intake, such as 3:10 PM, physiological data from the most recent five-minute time window of "15:05-15:10" is extracted. This yields an average cadence of 10 steps / minute and an average heart rate of 75 beats / minute during this period. Based on the correlation between these two values, a formula is used... The meaning and value determination process of each parameter in the formula for calculating metabolic fluctuation value are as follows: It is a metabolic fluctuation value used to quantify the degree of deviation of the current metabolic level from the baseline resting state. This is the average heart rate within the current time window, which is 75 beats per minute. It is the user's individualized resting heart rate, taken as the average heart rate of 58 beats per minute during the deep sleep period in the early morning of the same day. It is the average step rate of the current window, which is 10 steps per minute. It is the highest cadence recorded that day, which is 140 steps per minute, and the coefficient is... and These are weighting coefficients that reflect the contributions of heart rate changes and physical activity to overall metabolic rate, respectively. The weighting coefficients were set based on a cross-validation study using a double-labeled water method and an indirect energy metabolism measurement method. The rationale is that heart rate can sensitively and continuously reflect changes in comprehensive energy expenditure caused by multiple factors, including physical activity, emotions, and body temperature, making it a more comprehensive indicator of overall metabolic level. Therefore, the weights reflecting the contribution of heart rate are important. It was assigned a larger value of 0.7; while cadence mainly reflects limb activity and is one of the direct causes of metabolic changes, it serves as a supplementary and corrective term, thus reflecting the weight of cadence's contribution. Set the value to 0.3, and substitute all the values ​​into the formula for a complete calculation: Next, based on a study that correlates metabolic energy equivalents (METs) with... The research findings on correlation and calibration of calculated values ​​established a three-level identification interval for intraday metabolic status. When the calculated value is less than 0.15, it is determined to be a "resting metabolic state"; when... A value between 0.15 and 0.45 is considered a "low metabolic state"; when... A value greater than 0.45 is considered a "high metabolic state," and the calculated value will be used to determine this. The value 0.2266 was compared with the recognition interval, because Therefore, the user's current metabolic state is determined to be "low metabolic state", and a corresponding metabolic state classification label is assigned to the user.

[0027] The parameter screening and matching submodule filters nutrient absorption parameters corresponding to the current metabolic state from the enzyme activity parameters based on the enzyme activity regulation parameter group and metabolic state classification label, including pancreatic lipase absorption parameters, pancreatic amylase absorption parameters and trypsin absorption parameters, and outputs the absorption adaptation parameter group. The enzyme activity regulation parameter group {adjusted lipase parameter: 72.65, adjusted amylase parameter: 107.17, adjusted protease parameter: 59.81} and the "metabolic state classification label" "low metabolic state" are invoked. The enzyme activity parameters are then screened and further adjusted to generate nutrient absorption parameters most suitable for the current metabolic state. This process is achieved by introducing a set of preset nutrient utilization coefficients that strictly correspond to the metabolic state. The coefficients are set based on quantitative data from published physiological research literature regarding the human body's preference for the oxidation and utilization of three macronutrients under different activity intensities. The setting logic is as follows: in the "low metabolic state"... In a low metabolic state (equivalent to an activity intensity of 1.6-3.0 METs), the body prioritizes the use of carbohydrates, which are the easiest to break down, to meet immediate but not strenuous energy needs. Therefore, the utilization coefficient of carbohydrates is set to the highest value of 1.0. In contrast, the breakdown of proteins and lipids for energy is slower, and they are not the preferred energy source in this metabolic state, resulting in relatively lower utilization efficiency. Therefore, the utilization coefficients for both are set to 0.9. The utilization coefficients are then multiplied by the adjusted enzyme activity parameters to obtain the final parameters reflecting the current absorption capacity. Based on this, the pancreatic lipase absorption parameters are calculated as follows: The parameters for pancreatic amylase uptake were calculated as follows: The parameters for trypsin uptake were calculated as follows: The calculation results of the three metabolic state screening and matching are combined to output the absorption adaptability parameter group {pancreatic lipase absorption parameter: 65.39, pancreatic amylase absorption parameter: 107.17, trypsin absorption parameter: 53.83}.

[0028] Please see Figure 5 The conversion calculation module includes: The component parameter extraction submodule obtains the label information of the enteral nutrition preparation ingested by the user, extracts the labeled content values ​​of protein, carbohydrates and lipids per 100 ml, and generates a standard component content parameter set. We obtained the product label information of a specific enteral nutrition preparation (250 ml / bottle) that the user planned to consume at 3:15 PM. We carefully reviewed the nutrition facts label printed on the packaging in accordance with regulations, and accurately extracted and recorded the labeled content values ​​of the three major macronutrients per 100 ml of the preparation. Specifically, the extracted data were: protein content 8.0 g, carbohydrate content 20.0 g, and lipid content 7.0 g. We then structured and organized the three values ​​and corresponding units representing the standard formula of the nutritional preparation obtained from external authoritative information sources to generate a standard component content parameter set {protein content: 8.0 g / 100 ml, carbohydrate content: 20.0 g / 100 ml, lipid content: 7.0 g / 100 ml}. This parameter set will serve as the basic input data for calculating the actual amount of nutrients that the user can convert into nutrients in this intake.

[0029] The data call submodule matches the filtered nutrient absorption parameters in the absorption adaptability parameter group, establishes an association table with the corresponding three types of nutrients, and generates a corresponding group of absorption and conversion relationships. The absorption adaptation parameter set {pancreatic lipase absorption parameter: 65.39, pancreatic amylase absorption parameter: 107.17, trypsin absorption parameter: 53.83} is invoked, and a one-to-one, programmed association table is established with the three types of nutrients extracted from the nutritional supplement label. Specifically, the "lipids" nutrient in the label information is directly logically associated with the value of "pancreatic lipase absorption parameter" 65.39, the "carbohydrates" component is associated with the value of "pancreatic amylase absorption parameter" 107.17, and the "proteins" component is associated with the value of "trypsin absorption parameter" 53.83. The essence of this operation is to assign a quantified and personalized absorption capacity index to each macronutrient to be ingested, which has been adjusted for both the user's individual digestive system characteristics (gastric emptying) and the current overall metabolic state. This generates a clearly structured absorption-conversion correspondence set, which clarifies the pairing relationship between each nutrient and a highly individualized absorption efficiency parameter.

[0030] The transformation structure generation submodule transforms proteins, carbohydrates, and lipids according to the standard component content parameter group and the absorption and transformation correspondence group, including the protein transformation amount, carbohydrate transformation amount, and lipid transformation amount, and outputs nutrient transformation structure data; The standard ingredient content parameters {protein content: 8.0 g / 100 ml, carbohydrate content: 20.0 g / 100 ml, lipid content: 7.0 g / 100 ml} and the "absorption-conversion correspondence group" are used to estimate the absorbable and convertible amounts of protein, carbohydrates, and lipids in a planned 250 ml bottle of enteral nutrition formula. The conversion rate is calculated by comparing the user's individualized absorption parameters with the baseline parameters of a healthy population. This baseline parameter is based on statistical analysis of the corresponding enzyme activity data of two thousand healthy individuals, using the 50th percentile (median) of the distribution as the representative average level of the group. This provides an objective reference standard to assess the relative level of individual absorption capacity. Specifically, the baseline values ​​are pancreatic lipase 65.0, pancreatic amylase 110.0, and trypsin 55.0. The calculation formula is: Conversion rate = (Standard content / 100 ml) (intake volume) (Individual absorption parameter / benchmark parameter), therefore, protein conversion rate is grams, carbohydrate conversion rate is grams, lipid conversion rate is The final output is the nutritional conversion structure data for this intake: {protein conversion: 19.57g, carbohydrate conversion: 48.71g, lipid conversion: 17.60g}.

[0031] Please see Figure 6 The evaluation generation module includes: The intake record integration submodule obtains the time points of each user's enteral nutrient intake, arranges each time point in sequence, calls the nutrient conversion structure data corresponding to each time point, and generates an intake conversion time series data group. The system retrieved the user's two primary enteral nutrition intake times for the day, at 3:15 PM and 9:30 PM. These times were arranged chronologically, and the corresponding nutrient conversion data calculated for each time point was retrieved. The data for 3:15 PM was: {protein conversion: 19.57g, carbohydrate conversion: 48.71g, lipid conversion: 17.60g}. For the intake at 9:30 PM, due to changes in the user's physiological state, the user's metabolic state was reassessed and changed to "resting metabolic state". (Calculated values ​​less than 0.15) The corresponding nutrient conversion structure data were recalculated to obtain {protein conversion: 18.49 g, carbohydrate conversion: 38.97 g, lipid conversion: 19.56 g}. The two sets of data containing different time points and corresponding conversion amounts were integrated to generate a structured intake conversion time series data set that reflects dynamic changes [{time point: 15:15, conversion data: {protein: 19.57 g, carbohydrate: 48.71 g, lipid: 17.60 g}}, {time point: 21:30, conversion data: {protein: 18.49 g, carbohydrate: 38.97 g, lipid: 19.56 g}}].

[0032] The conversion difference calculation submodule calls the intake conversion time series data group, extracts the conversion amount of protein, carbohydrates and lipids between any two adjacent intakes, calculates the conversion difference respectively and classifies and combines the components to generate nutrient conversion difference combination data. Using the intake-conversion time series dataset, we extracted the conversion data of protein, carbohydrates, and lipids between two adjacent intake events at 3:15 PM and 9:30 PM. Then, we calculated the absolute difference in conversion between the later intake (9:30 PM) and the earlier intake (3:15 PM). The difference in protein conversion was 18.49 g - 19.57 g, with a calculated result of -1.08 g; the difference in carbohydrate conversion was 38.97 g - 48.71 g, with a calculated result of -9.74 g; and the difference in lipid conversion was 1 g. The difference between 9.56g and 17.60g resulted in a calculation of +1.96g. The three differences representing changes in conversion capacity were categorized and combined according to their respective nutrient components, ultimately generating combined data on nutrient conversion differences: {protein conversion difference: -1.08g, carbohydrate conversion difference: -9.74g, lipid conversion difference: +1.96g}. This data intuitively quantifies the specific dynamic changes in the absorption and conversion efficiency of various components in the same nutritional preparation between two intakes due to changes in the user's physiological state (mainly metabolic level).

[0033] The absorption status determination submodule judges the changing trends of protein, carbohydrate and lipid conversion based on the combination data of nutrient conversion differences, compares them with the preset absorption stability range, and outputs the enteral nutrition assessment results for the comprehensive judgment of enteral nutrition absorption efficiency, dynamic stability and component utilization. Based on the sign of the conversion difference values ​​for each nutrient in the nutrient conversion difference combination data, a trend judgment was first made. The protein conversion difference was -1.08 grams, indicating a downward trend in conversion; the carbohydrate conversion difference was -9.74 grams, also showing a downward trend; while the lipid conversion difference was +1.96 grams, showing an upward trend. Subsequently, the relative changes in the conversion amounts of each component were compared with a pre-defined absorption stability range. This range was set based on expert consensus in clinical nutritional support practice, using ±15% as the threshold. Fluctuations below this range are generally considered to be within the normal range of physiological regulation and have no significant clinical significance, while fluctuations exceeding this range may indicate malabsorption or metabolic adaptation problems, requiring attention. The specific comparison process was as follows: the relative change in protein was calculated... The absolute value of 5.5% is less than 15%, therefore the protein absorption state is determined to be stable; the relative change in lipids is calculated as follows: The absolute value of 11.1% is less than 15%, therefore the lipid absorption state is considered stable; the relative change in carbohydrates is calculated as follows: The absolute value of 20.0% is greater than 15%, therefore the carbohydrate absorption status is determined to be unstable. Based on the combined stability of all assessments, the following possible enteral nutrition evaluation results will be output according to the combination of the three nutrients: All Stable: If the absorption status of protein, lipids, and carbohydrates is determined to be stable, the output result will be: "Comprehensive Assessment: The user's absorption of all major macronutrients shows good dynamic stability, and the conversion rate fluctuates within an acceptable range with changes in physiological state, indicating that the current nutrition plan is well adapted to the user's physiological changes." Partially Unstable: If the absorption status of one or two nutrients is determined to be unstable, the output result will clearly indicate the specific unstable component, for example: "Comprehensive Assessment: The user's absorption of [stable nutrient] is stable." The absorption of nutrients showed good dynamic stability, but the dynamic stability for the absorption of [unstable nutrients] was insufficient, with conversion fluctuations exceeding the preset range. This indicates that the user's absorption adaptability to specific nutrients is poor, and attention should be paid to the supply and absorption of [unstable nutrients]. "All Unstable": If the absorption status of proteins, lipids, and carbohydrates is all determined to be unstable, the output result is: "Comprehensive Judgment: The user's absorption of all major macronutrients shows an unstable state, and the conversion rate varies with physiological state, exceeding the preset range. This suggests that the current nutritional plan may not be suitable for the user's physiological changes, or there may be potential absorption dysfunction. Further evaluation is recommended." In this example, since the absorption of protein and lipids is stable, while the absorption of carbohydrates is unstable, meeting the criteria of "partially unstable," the final assessment result is: "Overall judgment: The user's absorption of protein and lipids shows good dynamic stability, with the conversion amount fluctuating within an acceptable range as physiological state changes. However, the dynamic stability of carbohydrate absorption is insufficient, with the conversion amount showing a significant decrease of more than 15% after the metabolic level decreases, suggesting that the user's current carbohydrate utilization status may fluctuate considerably."

[0034] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A nutrition assessment system based on multi-source data, characterized in that, The system includes: The parameter extraction module acquires ultrasound images of the user's stomach at a specified time, analyzes the volume of gastric contents at each time, calculates the rate of descent of gastric contents over continuous time, calculates the total gastric emptying amplitude by calculating the volume difference between the initial and final gastric contents, and constructs a set of gastric emptying rate parameters. The parameter adjustment module acquires pancreatic digestive enzyme activity data from the user's fasting blood sample, and adjusts the weights of the pancreatic digestive enzyme activity data in conjunction with the gastric emptying rate parameter set to generate an enzyme activity adjustment parameter set. The parameter filtering module obtains the user's daily step frequency data and average heart rate, filters nutrient absorption parameters that match the metabolic state from the enzyme activity regulation parameter group, and outputs the absorption adaptation parameter group. The conversion calculation module obtains the protein, carbohydrate and lipid content per 100 ml of the enteral nutrition preparation ingested by the user, and calculates the conversion amount of the three types of parameters through the absorption adaptability parameter group to obtain nutrient conversion structure data. The assessment generation module acquires the nutrient conversion structure data after multiple enteral nutrition intakes by the user, calculates the difference in conversion amount in chronological order, judges the status of enteral nutrient absorption capacity, and obtains the enteral nutrition assessment result.

2. The nutrition assessment system based on multi-source data according to claim 1, characterized in that, The gastric emptying rate parameter set includes the rate of descent of gastric contents and the total amount of gastric contents emptied. The enzyme activity regulation parameter set includes adjusted enzyme activity values, digestive pathway correspondences, and component priority order. The absorption adaptation parameter set includes intraday metabolic status labels, intake behavior characteristics, and enzyme activity selection range. The nutrient conversion structure data includes protein conversion, carbohydrate conversion, and lipid conversion. The enteral nutrition assessment results include protein absorption trend status, carbohydrate absorption trend status, and lipid absorption trend status.

3. The nutrition assessment system based on multi-source data according to claim 1, characterized in that, The parameter extraction module includes: The image acquisition submodule acquires ultrasound images of the stomach at multiple time points when the user takes oral glucose solution, extracts the anteroposterior diameter boundary of the sagittal section of the gastric cavity in the image, and locates and marks the image by using image frame sequence and time label to generate anteroposterior diameter localization data. The volume analysis submodule, based on the anteroposterior diameter positioning data, calls the anteroposterior diameter change value in the image at each time point, combines the image shooting angle parameter and user body position information, and uses a geometric fitting method to establish the volume estimation relationship of the contents, generating a sequence of gastric contents volume values ​​at each time point; The emptying calculation submodule calls the gastric contents volume value sequence at each time point, calculates the gastric contents descent rate over a continuous period, and calculates the volume difference between the initial volume and the final volume as the total gastric contents emptying amplitude, generating a gastric emptying rate parameter set.

4. The nutrition assessment system based on multi-source data according to claim 3, characterized in that, The user was observed at multiple time points after taking oral glucose solution, namely 20 minutes, 40 minutes, and 60 minutes.

5. The nutrition assessment system based on multi-source data according to claim 3, characterized in that, The parameter adjustment module includes: The enzyme activity acquisition submodule acquires venous blood samples taken by the user in a fasting state, uses colorimetric method to measure the activity data of each type of pancreatic digestive enzyme, timestamps and identifies the measured values ​​with project numbers, and establishes a pancreatic digestive enzyme activity dataset. The parameter normalization submodule extracts the gastric contents descent rate and the total gastric contents emptying amplitude from the gastric emptying rate parameter group, performs normalization on the two parameters, and outputs normalized emptying parameter data. The weight adjustment submodule adjusts the weights of each type of pancreatic digestive enzyme activity data based on the pancreatic digestive enzyme activity dataset and the normalized emptying parameter data, respectively, to generate an enzyme activity regulation parameter set.

6. The nutrition assessment system based on multi-source data according to claim 5, characterized in that, The pancreatic digestive enzyme activity data include the activity values ​​of pancreatic lipase, pancreatic amylase, and trypsin.

7. The nutrition assessment system based on multi-source data according to claim 5, characterized in that, The parameter filtering module includes: The physiological parameter acquisition submodule acquires continuous cadence data recorded by the user's wearable device on the same day, calculates the range of cadence variation throughout the day, and combines the average heart rate values ​​during rest and activity periods. It then uses a time axis positioning method to synchronize and organize cadence and heart rate data, generating a combined daily exercise and heart rate data set. The metabolic state recognition submodule calls the combined daily exercise and heart rate data set, extracts the step frequency change amplitude and average heart rate value, calculates the metabolic fluctuation value based on the linkage relationship between the two values, and determines the current metabolic state based on the set daily metabolic state recognition interval, and assigns metabolic state classification labels. The parameter screening and matching submodule, based on the enzyme activity regulation parameter group and the metabolic state classification label, filters the nutrient absorption parameters corresponding to the current metabolic state from the enzyme activity parameters, including pancreatic lipase absorption parameters, pancreatic amylase absorption parameters and trypsin absorption parameters, and outputs the absorption adaptation parameter group.

8. The nutrition assessment system based on multi-source data according to claim 7, characterized in that, The conversion calculation module includes: The component parameter extraction submodule obtains the label information of the enteral nutrition preparation ingested by the user, extracts the labeled content values ​​of protein, carbohydrates and lipids per 100 ml, and generates a standard component content parameter set. The data absorption call submodule matches the screened nutrient absorption parameters in the absorption adaptability parameter group, establishes an association table with the corresponding three types of nutrients, and generates an absorption and conversion correspondence group. The transformation structure generation submodule, based on the standard component content parameter group and the absorption and transformation correspondence group, transforms proteins, carbohydrates and lipids respectively, including protein transformation amount, carbohydrate transformation amount and lipid transformation amount, and outputs nutrient transformation structure data.

9. The nutrition assessment system based on multi-source data according to claim 8, characterized in that, The evaluation generation module includes: The intake record integration submodule obtains the time points of each user's enteral nutrient intake, arranges each time point in sequence, calls the nutrient conversion structure data corresponding to each time point, and generates an intake conversion time series data group. The conversion difference calculation submodule calls the intake conversion time series data group, extracts the conversion amount of protein, carbohydrates and lipids between any two adjacent intakes, calculates the conversion amount difference respectively and performs component classification and combination to generate nutrient component conversion difference combination data. The absorption status determination submodule determines the changing trends of protein, carbohydrate and lipid conversion based on the nutrient conversion difference combination data, compares them with the preset absorption stability range, and outputs enteral nutrition assessment results for a comprehensive assessment of enteral nutrition absorption efficiency, dynamic stability and component utilization.

Images