Data analysis-based prognosis prediction method for patients with nutritional diseases

Abstract

The invention relates to the technical field of medical data analysis, in particular to a prognosis prediction method for a nutritional disease patient based on data analysis. The method comprises the following steps: collecting health monitoring data of a patient, including a diet log; reading and processing the diet log to obtain nutrition intervention data; calculating a real-time calorie intake difference value in a monitoring period based on the nutrition intervention data, judging the real-time calorie intake difference value in a calorie intake fluctuation curve according to a standard fluctuation range, and obtaining to-be-verified data based on a judgment result; determining a calorie intake variable coefficient based on the to-be-verified data and the nutrition prognosis index; a risk level is determined based on the calorie intake variation coefficient, the delay time of the nutritional intervention, and the weight change trajectory. According to the method, the average calorie intake variable coefficient is calculated based on the sliding window, abnormal diet records are detected, and repeated or contradictory nutrition intervention records are recognized and eliminated, so that the efficiency and accuracy of nutritional disease management are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of medical data analysis, and in particular to a method for predicting the prognosis of patients with nutritional diseases based on data analysis. Background Art

[0002] The core goal of prognosis prediction for patients with nutritional diseases is to analyze patient data to predict disease progression, treatment effects, or complication risks, thereby optimizing intervention strategies. Therefore, how to combine medical and health data, sensing technology, and remote monitoring technologies to provide more comprehensive and timely monitoring and intervention to ensure that patients can receive effective care and management after discharge, and how to combine digital medical technology to enhance the protective effect of prognosis prediction on brain health are increasingly becoming hot topics for technical personnel in related fields.

[0003] Chinese patent document CN113205880A discloses a method and device for predicting the prognosis of heart disease based on LogitBoost. The method includes: obtaining target CCTA clinical data, and determining multiple related variables from the target CCTA clinical data; binning the multiple related variables based on the multi-binning principle to obtain multiple binned variables corresponding to each of the related variables; calculating the variable information gain of the multiple binned variables corresponding to each of the related variables, and screening out several target related variables from the multiple related variables based on the variable information gain corresponding to all the related variables; establishing a LogitBoost prediction model based on the LogitBoost learning algorithm based on the target related variables and the target CCTA clinical data; It can be seen that the existing prognosis prediction method does not adequately handle the redundancy of dietary record data, which affects the efficiency of the model, and lacks associated pathological mechanisms and threshold adjustment, resulting in low model warning accuracy. Summary of the Invention

[0004] To this end, the present invention provides a prognosis prediction method for patients with nutritional diseases based on data analysis, so as to overcome the problem of low risk warning accuracy in the prior art due to the lack of multi-dimensional dynamic monitoring of patient data.

[0005] To achieve the above object, the present invention provides a method for predicting the prognosis of patients with nutritional diseases based on data analysis, comprising:

[0006] Collect patients' health monitoring data, including dietary logs, continuous blood glucose monitoring data, and exercise data;

[0007] Reading and processing the dietary log to obtain nutritional intervention data;

[0008] Calculating a real-time calorie intake difference within a monitoring period based on the nutritional intervention data, and drawing a calorie intake fluctuation curve, determining the real-time calorie intake difference in the calorie intake fluctuation curve according to a standard fluctuation range, and obtaining data to be verified based on the determination result;

[0009] Among them, the intake fluctuation curve is the curve of actual calorie intake changing over time;

[0010] Determine the coefficient of variation of caloric intake based on the data to be validated and nutritional prognostic indicators;

[0011] The corresponding risk level was determined based on the analysis results of the coefficient of variation of caloric intake, the delay time of nutritional intervention and the trajectory of weight change.

[0012] Further, calculating the real-time calorie intake difference within the monitoring period based on the nutritional intervention data includes:

[0013] Determine daily calorie recommendations based on the patient's basal metabolic rate and activity factor;

[0014] Obtaining actual calorie intake recorded in the nutritional intervention data;

[0015] Calculate the difference between actual calorie intake and daily recommended calorie intake to get the real-time calorie intake difference.

[0016] Furthermore, reading and processing the meal log includes:

[0017] Convert image data in the dietary log into text data, and convert audio data into text data;

[0018] Merge the converted data with the original text data to obtain the data to be scanned;

[0019] The data to be scanned is scanned and deleted to obtain nutritional intervention data.

[0020] Furthermore, scanning and deleting the data to be scanned includes:

[0021] Scanning the data to be scanned with a preset scanning length to obtain a current detection data segment;

[0022] Extracting real-time keywords from the current detection data segment, determining the number of the extracted real-time keywords, and obtaining keywords to be detected based on the determination result, or performing a matching determination on the next detection data segment;

[0023] Performing continuous extended scanning on the keyword to be detected to obtain an extended character segment;

[0024] The extended character segment is matched and detected according to the keyword to be detected, and whether to delete the extended character segment is determined according to the result of the matching detection to obtain the nutritional intervention data.

[0025] Furthermore, the data to be verified obtained based on the determination result includes:

[0026] Obtain each real-time calorie intake difference as the corresponding monitoring point;

[0027] Obtain the nutritional intervention data corresponding to the monitoring points that do not fall into the standard fluctuation range to obtain the data to be verified.

[0028] Furthermore, the coefficient of variation of caloric intake is determined based on the data to be verified and the nutritional prognostic indicators, including:

[0029] Obtain the standard deviation of actual calorie intake during the monitoring period, as well as the mean value of actual calorie intake during the monitoring period;

[0030] The coefficient of variation of calorie intake was obtained by calculating the quotient of the standard deviation of actual calorie intake and the mean of actual calorie intake.

[0031] Furthermore, risk levels were determined based on the coefficient of variation of caloric intake, the delay time of nutritional intervention, and the trajectory of weight change, including:

[0032] Compare the standard coefficient of variation with the coefficient of variation of caloric intake.

[0033] If the coefficient of variation of caloric intake is less than or equal to the standard coefficient of variation, inflammatory response indicators are monitored and the risk level is determined based on the monitoring results;

[0034] If the coefficient of variation of caloric intake is greater than the standard coefficient of variation, determine whether to adjust the standard coefficient of variation based on the weight change trajectory.

[0035] Further, determining whether to adjust the standard coefficient of variation based on the weight change trajectory includes:

[0036] Obtain the trajectory of actual calorie intake changes over time, and obtain the trajectory of weight changes after a preset delay time at any time;

[0037] Calculate the real-time similarity between the calorie intake change trajectory and the weight change trajectory;

[0038] Determine the real-time similarity based on the standard similarity.

[0039] If the real-time similarity is greater than or equal to the standard similarity, the inflammatory response indicators are monitored and the risk level is determined based on the monitoring results;

[0040] If the real-time similarity is less than the standard similarity, the standard coefficient of variation is adjusted to the corrected coefficient of variation;

[0041] The corrected coefficient of variation is the product of the standard coefficient of variation and the correction step size. The correction step size is the difference between 1 and the correction degree. The correction degree is the quotient of the part of the standard similarity that exceeds the real-time similarity and the standard similarity.

[0042] Furthermore, the inflammatory response indicators are monitored and the risk level is determined based on the monitoring results, including:

[0043] The inflammatory response indicators include C-reactive protein and carotid intima-media thickness;

[0044] Obtain the initial C-reactive protein level and initial carotid intima-media thickness, compare the standardized C-reactive protein level with the initial C-reactive protein level, and compare the standardized carotid intima-media thickness with the initial carotid intima-media thickness,

[0045] If the initial C-reactive protein level is lower than the standard C-reactive protein level, and the initial carotid intima-media thickness is lower than the standard carotid intima-media thickness, the current risk level is determined to be level one;

[0046] If the initial C-reactive protein level is greater than or equal to the standard C-reactive protein level, and the initial carotid intima-media thickness is greater than or equal to the standard carotid intima-media thickness, the current risk level is determined to be level 4 risk;

[0047] If the initial C-reactive protein level is greater than or equal to the standard C-reactive protein level, or the initial carotid intima-media thickness is greater than or equal to the standard carotid intima-media thickness, the current disease stage is obtained, the type of the current disease stage is analyzed, and the current risk level is determined based on the analysis results and the real-time slope of the thickness change curve.

[0048] Further, determining the current risk level based on the analysis results and the real-time slope of the thickness change curve includes:

[0049] If the current disease stage is stage 1, the current risk level is determined to be level 2 risk;

[0050] If the current disease stage is the second stage, the thickness change curve is obtained, the C-reactive protein change curve is monitored, and the real-time slope of the thickness change curve is monitored.

[0051] If the real-time slope is less than or equal to the standard slope, the current risk level is determined to be level three risk;

[0052] If the real-time slope is greater than the standard slope, the current risk level is determined to be level four risk.

[0053] Compared with the existing technology, the beneficial effect of the present invention lies in that it screens indicators that are strongly correlated with nutritional prognosis, such as the HOMA-I R index and the inflammatory marker CRP, and calculates the average calorie intake coefficient of variation in the past 30 days based on a sliding window, detects abnormal dietary records, identifies and eliminates duplicate or contradictory nutritional intervention records, improves the efficiency and accuracy of nutritional disease management, and achieves accurate and dynamic prognosis prediction through data-driven methods. It can be adapted to various nutrition-related diseases such as diabetes, osteoporosis, and chronic liver disease, and provide quantitative support for personalized nutritional therapy.

[0054] Furthermore, by analyzing and processing the monitored data, repeated content appearing in the text data can be efficiently identified, that is, by performing extended scanning detection on the keywords, and continuously obtaining extended character segments adjacent to the keyword and with the same character length, until it is determined that the extended character segment does not match the keyword, the extended scanning detection of the current keyword is stopped, thereby avoiding repeated matching of repeated real-time keywords in the database, that is, avoiding repeated detection of matches, improving data processing efficiency, and deleting repeated content to process data in a timely and effective manner, reducing repeated and redundant information in nutritional data, and improving data quality and availability. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 Schematic diagram of a flow chart of a method for predicting prognosis of patients with nutritional diseases based on data analysis according to an embodiment of the present invention;

[0056] Figure 2 A schematic diagram of a process for obtaining nutritional intervention data according to an embodiment of the present invention;

[0057] Figure 3 A schematic diagram of a process for determining the coefficient of variation of caloric intake according to an embodiment of the present invention;

[0058] Figure 4 Schematic diagram of a flow chart for determining whether to adjust the standard coefficient of variation based on a weight change trajectory according to an embodiment of the present invention. DETAILED DESCRIPTION

[0059] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are merely used to explain the present invention and are not intended to limit the present invention.

[0060] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0061] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside", and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.

[0062] Furthermore, it should be noted that, in the description of the present invention, unless otherwise expressly specified or limited, the terms "mounted," "connected," and "connected" should be understood in a broad sense. For example, they may refer to fixed connections, detachable connections, or integral connections; mechanical connections or electrical connections; direct connections or indirect connections through an intermediate medium; and internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0063] See also Figure 1 As shown, it is a flow chart of a method for predicting the prognosis of patients with nutritional diseases based on data analysis according to an embodiment of the present invention. The present invention provides a method for predicting the prognosis of patients with nutritional diseases based on data analysis, comprising:

[0064] Step S1, collecting the patient's health monitoring data, including meal logs, continuous blood glucose monitoring data, and exercise data;

[0065] Step S2, reading and processing the dietary log to obtain nutritional intervention data;

[0066] Step S3, calculating the real-time calorie intake difference within the monitoring period based on the nutritional intervention data, and drawing a calorie intake fluctuation curve, determining the real-time calorie intake difference in the calorie intake fluctuation curve according to a standard fluctuation range, and obtaining data to be verified based on the determination result;

[0067] Among them, the intake fluctuation curve is the curve of actual calorie intake changing over time;

[0068] Step S4, determining the coefficient of variation of calorie intake based on the data to be verified and the nutritional prognosis indicator;

[0069] Step S5: determining the corresponding risk level based on the analysis results of the coefficient of variation of caloric intake, the delay time of nutritional intervention, and the weight change trajectory.

[0070] In this example, clinical data including age, BMI, medical history, laboratory indicators (human serum albumin, HbA1c, vitamin D level), nutritional data including dietary log (recorded via app), body composition analysis (bioelectrical impedance), metabolomics data (short-chain fatty acid profile), and time series data including continuous glucose monitoring (CGM) and exercise volume (wearable device) are collected. The FHIR standard is used to integrate electronic health record (EHR) and wearable device data. The OMOP-CDM model is used to map heterogeneous data to ensure cross-platform compatibility. Indicators strongly related to nutritional prognosis, such as HOMA-I, are screened. R index, inflammatory marker CRP, and the coefficient of variation of average calorie intake over the past 30 days calculated based on a sliding window detect abnormal dietary records, identify and eliminate duplicate or contradictory nutritional intervention records, improve the efficiency and accuracy of nutritional disease management, and achieve accurate and dynamic prognosis prediction through data-driven methods. It can be adapted to various nutrition-related diseases such as diabetes, osteoporosis, and chronic liver disease, providing quantitative support for personalized nutritional treatment. The prognosis prediction of patients with nutritional diseases is achieved through multi-dimensional management and control, achieving timely intervention, precise management, and comprehensive protection of brain health. For example, long-term diabetes increases the risk of cerebral arteriosclerosis, leading to stroke or vascular dementia, and may also cause peripheral neuropathy and diabetic encephalopathy. It may also cause temporary brain function disorders due to acute complications such as hypoglycemia or ketoacidosis. Therefore, by monitoring and analyzing the prognostic data of patients with the disease and accurately assessing the disease progression and complication risks, it is helpful to intervene early, delay the course of the disease, and improve the efficiency of prognosis management.

[0071] Specifically, calculating the real-time calorie intake difference during the monitoring period based on the nutritional intervention data includes:

[0072] Determine daily calorie recommendations based on the patient's basal metabolic rate and activity factor;

[0073] Obtaining actual calorie intake recorded in the nutritional intervention data;

[0074] Calculate the difference between actual calorie intake and daily recommended calorie intake to get the real-time calorie intake difference.

[0075] In this embodiment, the basal metabolic rate is calculated according to the patient's gender, age, height and activity level, and the activity coefficient is determined according to Harris-Benedict. The daily recommended calories are the product of the basal metabolic rate and the activity coefficient; for example, the patient's gender is male, age: 30 years old, height: 175 cm, weight: 70 kg, activity level is moderate activity, and the activity coefficient corresponding to moderate activity is 1.55, then the basal metabolic rate is 10×weight + 6.25×height-5×age+5=1648.75; daily recommended calories=1648.75×1.55=2550 kcal / day; obtaining the actual calorie intake recorded in the dietary log is to identify data related to meal ingredients in text data, including using OCR technology (Tesseract engine) to extract text information from images, and applying NLP model (BERT-Base) to parse diet descriptions in voice logs (such as "I ate two slices of whole wheat bread for breakfast"), and docking with nutrition database (USDA FoodData Center l) Map the food names to calories, protein, fat, carbohydrates and other ingredients, obtain the sum of the calories corresponding to each food, and get the actual calorie intake.

[0076] See Figure 2 As shown, it is a schematic diagram of the process of obtaining nutritional intervention data according to an embodiment of the present invention;

[0077] Specifically, reading and processing the meal log includes:

[0078] Step S2001, converting the image data in the meal log into text data, and converting the audio data into text data;

[0079] Step S2002: merging the converted data with the original text data to obtain data to be scanned;

[0080] Step S2003: Scan and delete the data to be scanned to obtain nutritional intervention data.

[0081] The dietary log in this embodiment can be voice data or text data. Reading and processing the dietary log includes converting the image data in the dietary log into text data, converting the audio data into text data, and then merging it with the original text data to obtain nutritional intervention data.

[0082] Specifically, scanning and deleting the data to be scanned includes:

[0083] Scanning the data to be scanned with a preset scanning length to obtain a current detection data segment;

[0084] Extracting real-time keywords from the current detection data segment, determining the number of the extracted real-time keywords, and obtaining keywords to be detected based on the determination result, or performing a matching determination on the next detection data segment;

[0085] Performing continuous extended scanning on the keyword to be detected to obtain an extended character segment;

[0086] The extended character segment is matched and detected according to the keyword to be detected, and whether to delete the extended character segment is determined according to the result of the matching detection to obtain the nutritional intervention data.

[0087] By analyzing and processing the monitored data, repeated content in the text data can be efficiently identified, that is, by performing extended scanning detection on the keywords, and continuously obtaining extended character segments adjacent to the keyword and with the same character length until it is determined that the extended character segment does not match the keyword, the extended scanning detection of the current keyword is stopped, thereby avoiding repeated matching of repeated real-time keywords in the database, that is, avoiding repeated detection of matches, improving data processing efficiency, and deleting repeated content to process data in a timely and effective manner, reducing repeated and redundant information in nutritional data, and improving data quality and availability.

[0088] Specifically, obtaining a keyword to be detected based on the determination result, or performing a matching determination on the next detection data segment includes:

[0089] If the number of the extracted real-time keywords is greater than or equal to one, a real-time keyword is selected in chronological order, and a matching determination is performed on the real-time keywords to obtain a keyword to be detected or an unassociated data segment;

[0090] If the number of the extracted real-time keywords is less than one, performing a matching determination on the next detection data segment;

[0091] The preset scanning length in this embodiment is the size of the scanning window, which is set by the user according to needs and can be a fixed length of characters or words. The sliding window starts from the starting position of the text data and slides on the text data with a set length, with a fixed step size each time. The set preset scanning length is adjustable to flexibly scan and process the text data to adapt to different text processing needs; the real-time keywords are nutrition-related keywords, and the real-time keywords are extracted from the nutrition vocabulary by matching the words in the text data with the nutrition-related words in the vocabulary to extract keywords.

[0092] By obtaining the current detection data segment, real-time keywords in the current detection data segment are extracted to process or analyze the text data in the sliding window at each sliding window position. By adopting a fast scanning method, the input content is analyzed and processed. The method is simple and effective.

[0093] Specifically, the real-time keywords are matched and determined to obtain keywords to be detected, or unassociated data segments include:

[0094] Matching the real-time keywords with the nutrition keywords in the nutrition database,

[0095] If there is a match with the nutrition keyword, the keyword to be detected is obtained;

[0096] If there is no match with the nutritional keyword, continue to match the next real-time keyword until the matching judgment of each real-time keyword in the current detection data segment is completed. If it is determined that there is no match with the nutritional keyword, the current detection data segment is marked as an unassociated data segment.

[0097] By matching the extracted keywords in chronological order when it is determined that the number of the extracted real-time keywords is greater than or equal to one, the orderliness of data processing is guaranteed. Since the data input by the patient may be incomplete or the content is irrelevant to the condition, the data needs to be analyzed in detail to ensure the accuracy of the data in a timely manner. At the same time, after completing the scanning analysis, if it is still determined that there is no match with the nutritional keyword, it means that the content is irrelevant to the condition. Through timely warning and intervention, the patient's abnormal condition is discovered in time, thereby improving the patient's monitoring and management effect.

[0098] Specifically, the continuous extended scan includes:

[0099] The character length of the keyword to be detected is used as the detection character length;

[0100] Taking the next adjacent character of the keyword to be detected as the starting node;

[0101] The text data is continuously extended and scanned with the detected character length to obtain an extended character segment.

[0102] In this embodiment, since consecutive words may appear repeatedly, after matching a keyword, the keyword is subjected to extended scanning detection. By continuously obtaining extended character segments adjacent to the keyword and having the same character length, the extended scanning detection of the current keyword is stopped until it is determined that the extended character segment does not match the keyword, thereby avoiding repeated matching of repeated real-time keywords in the database, that is, avoiding repeated detection of matches, and improving data processing efficiency.

[0103] Specifically, any of the nutrition keywords corresponds to an associated keyword corresponding to the nutrition keyword;

[0104] The result of the matching detection includes obtaining the nutrition keyword and the associated keyword corresponding to the keyword to be detected in the nutrition database as a plurality of keywords to be matched, and performing matching detection on the extended character segment according to each of the keywords to be matched.

[0105] If the keyword to be matched matches the extended character segment, it is determined that duplicate data exists, the extended character segment is deleted, and the keyword to be detected is continuously extended and scanned to obtain the extended character segment, so as to perform matching detection on the extended character segment;

[0106] If the keyword to be matched does not match the extended character segment, it is determined that no duplicate data occurs, and the continuous extended scan of the keyword to be detected is completed. The next keyword to be detected is continuously extended and scanned to obtain the corresponding extended character segment to determine whether to delete the extended character segment.

[0107] In this embodiment, since the multiple expressions of food names also cause data duplication and redundancy, multiple expressions of the same food are stored in the nutrition database to facilitate timely and effective data processing during keyword matching analysis, reduce duplication and redundant information in nutritional intervention data, and improve data quality and availability.

[0108] Specifically, the data to be verified based on the determination result includes:

[0109] Obtain each real-time calorie intake difference as the corresponding monitoring point;

[0110] Obtain the nutritional intervention data corresponding to the monitoring points that do not fall into the standard fluctuation range to obtain the data to be verified.

[0111] See Figure 3 , which is a schematic diagram of a process for determining the coefficient of variation of caloric intake according to an embodiment of the present invention;

[0112] Specifically, the coefficient of variation of caloric intake was determined based on the data to be verified and the nutritional prognostic indicators, including:

[0113] Step S401, obtaining the standard deviation of actual calorie intake during the monitoring period and the mean value of actual calorie intake during the monitoring period;

[0114] Step S402, calculating the quotient of the standard deviation of actual calorie intake and the mean of actual calorie intake to obtain the coefficient of variation of calorie intake.

[0115] In this embodiment, the standard deviation of actual calorie intake represents the fluctuation of actual calorie intake, and the coefficient of variation of calorie intake represents the degree of fluctuation of actual calorie intake. The larger the coefficient of variation of calorie intake, the greater the degree of fluctuation of actual calorie intake.

[0116] Specifically, risk levels based on the coefficient of variation of caloric intake, delay in nutritional intervention, and trajectory of weight change include:

[0117] Compare the standard coefficient of variation with the coefficient of variation of caloric intake.

[0118] If the coefficient of variation of caloric intake is less than or equal to the standard coefficient of variation, inflammatory response indicators are monitored and the risk level is determined based on the monitoring results;

[0119] If the coefficient of variation of caloric intake is greater than the standard coefficient of variation, determine whether to adjust the standard coefficient of variation based on the weight change trajectory.

[0120] See Figure 4 As shown, it is a flow chart of determining whether to adjust the standard coefficient of variation based on the weight change trajectory according to an embodiment of the present invention;

[0121] Specifically, whether to adjust the standardized coefficient of variation based on the weight change trajectory includes:

[0122] Step S501, obtaining a trajectory of calorie intake changes over time, and obtaining a trajectory of weight changes after a preset delay time at any time;

[0123] Step S502, calculating the real-time similarity between the calorie intake change trajectory and the weight change trajectory;

[0124] Step S503: Determine the real-time similarity based on the standard similarity.

[0125] If the real-time similarity is greater than or equal to the standard similarity, the inflammatory response indicators are monitored and the risk level is determined based on the monitoring results;

[0126] If the real-time similarity is less than the standard similarity, the standard coefficient of variation is adjusted to the corrected coefficient of variation;

[0127] The corrected coefficient of variation is the product of the standard coefficient of variation and the correction step size. The correction step size is the difference between 1 and the correction degree. The correction degree is the quotient of the part of the standard similarity that exceeds the real-time similarity and the standard similarity.

[0128] The preset delay time in this embodiment represents a time interval between the calorie intake and weight changes of patients with nutritional diseases, which is set to 7 to 14 days. The set value is related to the type of disease. For example, it is set to 10 days for diabetic patients. The standard similarity is the time series matching threshold of the calorie intake change and the weight change, which is used to determine whether the calorie fluctuation directly leads to abnormal weight. The standard similarity is set to 0.8. If the real-time similarity is greater than or equal to the standard similarity, it is considered that the calorie fluctuation is highly correlated with the weight change. If the real-time similarity is less than the standard similarity, it means that the correlation between the calorie fluctuation and the weight change is not high, which may be caused by the disease. In this way, the standard coefficient of variation is reduced to monitor the inflammatory response indicators, improve the sensitivity to calorie fluctuations, and accurately identify the risk level.

[0129] Specifically, the inflammatory response indicators are monitored, and the risk level is determined based on the monitoring results, including:

[0130] The inflammatory response indicators include C-reactive protein and carotid intima-media thickness;

[0131] Obtain the initial C-reactive protein level and initial carotid intima-media thickness, compare the standardized C-reactive protein level with the initial C-reactive protein level, and compare the standardized carotid intima-media thickness with the initial carotid intima-media thickness,

[0132] If the initial C-reactive protein level is lower than the standard C-reactive protein level, and the initial carotid intima-media thickness is lower than the standard carotid intima-media thickness, the current risk level is determined to be level one;

[0133] If the initial C-reactive protein level is greater than or equal to the standard C-reactive protein level, and the initial carotid intima-media thickness is greater than or equal to the standard carotid intima-media thickness, the current risk level is determined to be level 4 risk;

[0134] If the initial C-reactive protein level is greater than or equal to the standard C-reactive protein level, or the initial carotid intima-media thickness is greater than or equal to the standard carotid intima-media thickness, the current disease stage is obtained and the type of the current disease stage is analyzed.

[0135] If the current disease stage is stage 1, the current risk level is determined to be level 2 risk;

[0136] If the current disease stage is the second stage, the thickness change curve is obtained, the C-reactive protein change curve is monitored, and the real-time slope of the thickness change curve is monitored.

[0137] If the real-time slope is less than or equal to the standard slope, the current risk level is determined to be level three risk;

[0138] If the real-time slope is greater than the standard slope, the current risk level is determined to be level 4 risk;

[0139] In this example, the first-level risk is stable acute risk; the second-level risk is the risk of possible recurrence, which requires lifestyle adjustments, such as a low-GI diet and regular exercise, to delay vascular structural damage; the third-level risk is recurrence, which requires anti-infection treatment, such as SGLT2 inhibitors and combined lipid-lowering drugs (statins);

[0140] Level 4 risk is an extremely high risk of developing diabetic retinopathy, which requires emergency intervention; the first stage is the early stage of diabetes, and the second stage is the middle and late stages of diabetes. The early and middle and late stages of diabetes are related to HbA1c and the duration of the disease. The first stage (early stage): HbA1c is less than or equal to 7.5%, the course of the disease is less than 5 years, and there are no complications. The second stage (middle and late stages): HbA1c is greater than 7.5%, the course of the disease is greater than or equal to 5 years, and there is retinopathy; the standard slope is set at 0.1mm / year.

[0141] Since the progression of retinopathy indirectly reflects the health status of brain microvessels, stroke or vascular dementia, diabetic encephalopathy, and temporary brain dysfunction may occur. Therefore, intelligent risk stratification is used to predict the risk of retinopathy through big data analysis, timely remind medical intervention, and achieve accurate management of prognosis data.

[0142] Thus far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art may make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will fall within the scope of protection of the present invention.

[0143] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that the present invention is susceptible to various modifications and variations. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. A method for predicting the prognosis of patients with nutritional diseases based on data analysis, characterized in that: include, Collect patients' health monitoring data, including dietary logs, continuous blood glucose monitoring data, and exercise data; Reading and processing the dietary log to obtain nutritional intervention data; The real-time calorie intake difference within the monitoring period is calculated based on the nutritional intervention data, and a calorie intake fluctuation curve is drawn. The real-time calorie intake difference in the calorie intake fluctuation curve is determined according to a standard fluctuation range, and data to be verified is obtained based on the determination result, wherein: The intake fluctuation curve is the curve of actual calorie intake changing over time; determining a coefficient of variation of caloric intake based on the data to be verified and the nutritional prognostic indicator; The corresponding risk level was determined based on the analysis results of the coefficient of variation of caloric intake, the delay time of nutritional intervention and the trajectory of weight change.

2. The method for predicting the prognosis of a patient with a nutritional disease based on data analysis according to claim 1, wherein: Calculating the real-time calorie intake difference during the monitoring period based on the nutritional intervention data includes: Determine daily calorie recommendations based on the patient's basal metabolic rate and activity factor; Obtaining actual calorie intake recorded in the nutritional intervention data; Calculate the difference between actual calorie intake and daily recommended calorie intake to get the real-time calorie intake difference.

3. The method for predicting the prognosis of a patient with a nutritional disease based on data analysis according to claim 1, wherein: Reading and processing the meal log includes: Convert image data in the dietary log into text data, and convert audio data into text data; Merge the converted data with the original text data to obtain the data to be scanned; The data to be scanned is scanned and deleted to obtain nutritional intervention data.

4. The method for predicting the prognosis of a patient with a nutritional disease based on data analysis according to claim 3, characterized in that: Scanning and deleting the data to be scanned includes: Scanning the data to be scanned with a preset scanning length to obtain a current detection data segment; Extracting real-time keywords from the current detection data segment, determining the number of the extracted real-time keywords, and obtaining keywords to be detected based on the determination result, or performing a matching determination on the next detection data segment; Performing continuous extended scanning on the keyword to be detected to obtain an extended character segment; The extended character segment is matched and detected according to the keyword to be detected, and whether to delete the extended character segment is determined according to the result of the matching detection to obtain the nutritional intervention data.

5. The method for predicting the prognosis of a patient with a nutritional disease based on data analysis according to claim 1, wherein: The data to be verified based on the judgment result includes: Obtain each real-time calorie intake difference as the corresponding monitoring point; Obtain the nutritional intervention data corresponding to the monitoring points that do not fall into the standard fluctuation range to obtain the data to be verified.

6. The method for predicting the prognosis of a patient with a nutritional disease based on data analysis according to claim 5, characterized in that: The coefficient of variation of caloric intake was determined based on the data to be validated and nutritional prognostic indicators, including: Obtain the standard deviation of actual calorie intake during the monitoring period, as well as the mean value of actual calorie intake during the monitoring period; The coefficient of variation of calorie intake was obtained by calculating the quotient of the standard deviation of actual calorie intake and the mean of actual calorie intake.

7. The method for predicting the prognosis of patients with nutritional diseases based on data analysis according to claim 6, characterized in that: Risk levels based on coefficient of variation of caloric intake, delay in nutritional intervention, and trajectory of weight change include: Compare the standard coefficient of variation with the coefficient of variation of caloric intake. If the coefficient of variation of caloric intake is less than or equal to the standard coefficient of variation, inflammatory response indicators are monitored and the risk level is determined based on the monitoring results; If the coefficient of variation of caloric intake is greater than the standard coefficient of variation, determine whether to adjust the standard coefficient of variation based on the weight change trajectory.

8. The method for predicting the prognosis of nutritional disease patients based on data analysis according to claim 7, characterized in that: Determining whether to adjust the standardized coefficient of variation based on the weight change trajectory includes: Obtain the trajectory of actual calorie intake changes over time, and obtain the trajectory of weight changes after a preset delay time at any time; Calculate the real-time similarity between the calorie intake change trajectory and the weight change trajectory; Determine the real-time similarity based on the standard similarity. If the real-time similarity is greater than or equal to the standard similarity, the inflammatory response indicators are monitored and the risk level is determined based on the monitoring results; If the real-time similarity is less than the standard similarity, the standard coefficient of variation is adjusted to the corrected coefficient of variation; The corrected coefficient of variation is the product of the standard coefficient of variation and the correction step size. The correction step size is the difference between 1 and the correction degree. The correction degree is the quotient of the part of the standard similarity that exceeds the real-time similarity and the standard similarity.

9. The method for predicting the prognosis of nutritional disease patients based on data analysis according to claim 7, characterized in that: Monitor inflammatory response indicators and determine risk levels based on the monitoring results, including: The inflammatory response indicators include C-reactive protein and carotid intima-media thickness; Obtain the initial C-reactive protein level and initial carotid intima-media thickness, compare the standardized C-reactive protein level with the initial C-reactive protein level, and compare the standardized carotid intima-media thickness with the initial carotid intima-media thickness, If the initial C-reactive protein level is lower than the standard C-reactive protein level, and the initial carotid intima-media thickness is lower than the standard carotid intima-media thickness, the current risk level is determined to be level one; If the initial C-reactive protein level is greater than or equal to the standard C-reactive protein level, and the initial carotid intima-media thickness is greater than or equal to the standard carotid intima-media thickness, the current risk level is determined to be level 4 risk; If the initial C-reactive protein level is greater than or equal to the standard C-reactive protein level, or the initial carotid intima-media thickness is greater than or equal to the standard carotid intima-media thickness, the current disease stage is obtained, the type of the current disease stage is analyzed, and the current risk level is determined based on the analysis results and the real-time slope of the thickness change curve.

10. The method for predicting the prognosis of nutritional disease patients based on data analysis according to claim 9, characterized in that: Determine the current risk level based on the analysis results and the real-time slope of the thickness change curve, including: If the current disease stage is stage 1, the current risk level is determined to be level 2 risk; If the current disease stage is the second stage, the thickness change curve is obtained, the C-reactive protein change curve is monitored, and the real-time slope of the thickness change curve is monitored. If the real-time slope is less than or equal to the standard slope, the current risk level is determined to be level three risk; If the real-time slope is greater than the standard slope, the current risk level is determined to be level four risk.

Images