Diabetes risk assessment method and system

Abstract

The invention relates to the technical field of diabetes risk assessment, and discloses a diabetes risk assessment method and system, and the method comprises the steps: collecting user data, obtaining a diabetes diagnosis standard, and carrying out the judgment of a crowd according to the user data and the diabetes diagnosis standard; establishing a life correction model, a saccharification deviation model and a glycosylation accumulation model, and processing the user data to obtain analysis data; when the crowd is judged to be a non-diabetic crowd, establishing a disease assessment model, calculating a basic risk score by the disease assessment model based on a life correction model, auditing and correcting the basic risk score through a disease condition to obtain a glycosylation correction risk score, and mapping the glycosylation correction risk score into a disease risk level; and when the crowd is determined to be the diabetic crowd, establishing a control evaluation model, calculating a basic control risk level by the control evaluation model based on the life correction model and the glycosylation accumulation model, and auditing and correcting the basic control risk level through a control condition to obtain the control risk level.

Description

Technical Field

[0001] This invention relates to the field of diabetes risk assessment technology, specifically a diabetes risk assessment method and system. Background Technology

[0002] Diabetes mellitus is a metabolic disease characterized by chronic hyperglycemia, with an insidious and long-term onset. Current risk assessment methods for diabetes often employ single-population approaches, either predicting risk separately for non-diabetic individuals or analyzing glycemic control in diabetic patients. However, in actual health management, two population groups often coexist within the same system: individuals with potential risk but not yet diagnosed, and diagnosed and treated diabetic patients. Existing methods cannot simultaneously assess the health risks of both population groups within a unified framework, leading to fragmented assessment results, poor data continuity, and difficulty in maintaining consistency over long-term health monitoring.

[0003] Meanwhile, traditional risk assessment methods generally rely on single-point blood glucose or single-sample glycated hemoglobin (HbA1c) tests, failing to consider key factors such as blood glucose fluctuations, individual differences in HbA1c levels, and long-term exposure to hyperglycemia. This results in significant biases between individuals and makes it difficult to accurately reflect the true risk level. Furthermore, current assessments of diabetes control primarily rely on static indicators (such as HbA1c target achievement rates), lacking dynamic analysis mechanisms that combine blood glucose time-series data with HbA1c physiological characteristics, making it difficult to achieve individualized quantitative risk assessment.

[0004] In summary, there is an urgent need for a unified method that can simultaneously identify population types and separately assess the risk of disease and the risk of disease control, so as to achieve continuous and quantifiable risk assessment for both non-diabetic and diabetic populations. Summary of the Invention

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a method for assessing diabetes risk, comprising: collecting user data, obtaining diabetes diagnostic criteria, and performing population determination based on user data and diabetes diagnostic criteria; Establish lifetime correction models, glycosylation deviation models, and glycosylation accumulation models; process user data, and obtain analytical data. When the population is determined to be non-diabetic, an incidence assessment model is established. The incidence assessment model calculates the basic risk score based on the lifespan correction model. The basic risk score is corrected by the audit of the incidence conditions to obtain the glycosylation-corrected risk score. The glycosylation-corrected risk score is then mapped to the incidence risk level. When a population is identified as having diabetes, a control assessment model is established. The control assessment model calculates the basic control risk level based on the lifespan correction model and the glycosylation accumulation model. The basic control risk level is then corrected through a control condition audit to obtain the control risk level.

[0007] As a preferred embodiment of the diabetes risk assessment method described in this invention, the step of collecting user data, obtaining diabetes diagnostic criteria, and performing population determination based on user data and diabetes diagnostic criteria includes the user data including glycated hemoglobin data, fasting blood glucose data, and blood glucose time series data. The stated diagnostic criteria for diabetes are numerical thresholds for diagnosing diabetes. User data is compared with these diagnostic criteria to determine the target population. When fasting blood glucose levels are ≥7.0 mmol / L or glycated hemoglobin levels are ≥6.5%, the population is classified as having diabetes. When fasting blood glucose is <7.0 mmol / L and glycated hemoglobin is <6.5%, the population is considered non-diabetic.

[0008] As a preferred embodiment of the diabetes risk assessment method described in this invention, the establishment of a lifespan correction model includes: calculating the glycation rate coefficient based on average blood glucose, calculating the equivalent red blood cell lifespan based on glycated hemoglobin data and the glycation rate coefficient, correcting the glycated hemoglobin data based on the equivalent red blood cell lifespan, and outputting the corrected glycated hemoglobin data. Establishing a glycated hemoglobin deviation model involves calculating the expected glycated hemoglobin data based on the average blood glucose level, subtracting the glycated hemoglobin data from the glycated hemoglobin data, and outputting the glycated deviation value. The establishment of the glycosylation accumulation model includes setting the hyperglycemia threshold to 7.8 mmol / L based on the diagnostic criteria for diabetes, and calculating the excess exposure integral based on blood glucose time series data: ; Where E represents the over-threshold exposure integral; N represents the total number of blood glucose time series data; i represents the i-th data point in the blood glucose time series data; max represents taking the maximum value; This represents the blood glucose value in the i-th blood glucose time series data; This represents the blood glucose value in the (i+1)th blood glucose time series data; express and Time difference; The over-threshold exposure integral is normalized over the entire duration to calculate the average over-threshold exposure intensity. The average over-threshold exposure intensity and glycated hemoglobin data are linearly synthesized to obtain the cumulative amount of glycosylation. The data analyzed included corrected glycated hemoglobin data, glycation deviation values, and cumulative glycation.

[0009] As a preferred embodiment of the diabetes risk assessment method described in this invention, the disease assessment model calculates a basic risk score based on a lifespan correction model, corrects the basic risk score through disease condition auditing, obtains a glycosylation-corrected risk score, and maps the glycosylation-corrected risk score to a disease risk level, including obtaining corrected glycosylated hemoglobin data according to the lifespan correction model and obtaining a glycosylation deviation value according to the glycosylation deviation model. Using the corrected glycated hemoglobin data as the basic input of the disease assessment model, the Sigmoid function is used to map the corrected glycated hemoglobin data to a range of 0 to 1 to obtain the basic risk score; The baseline risk score was adjusted by auditing the conditions for the onset of the disease. The basic risk score is adjusted based on the saccharification deviation value. If the absolute value of the saccharification deviation value is greater than or equal to 0.6%, the saccharification deviation value is called, and the basic risk score is adjusted by an exponential function to output the saccharification-corrected risk score. If the absolute value of the saccharification deviation value is less than 0.6%, no adjustment is made, and the saccharification-corrected risk score is equal to the basic risk score. The glycation correction risk score is adjusted based on the cumulative amount of glycosylation. When any glycosylation correction condition is met, the cumulative amount of glycosylation is compressed using the Sigmoid function to obtain an intensity score. The glycation correction risk score is calculated based on the glycation correction risk score. The glycation correction risk score is calculated and output based on the glycation correction coefficient and the intensity score. If the glycosylation correction condition is not met, no correction is performed, and the glycation correction risk score is equal to the glycation correction risk score. The glycosylation correction conditions include two: Blood glucose time series data with a continuous duration of greater than 7.8 mmol / L for 30 minutes or more; The cumulative duration of blood glucose time series data greater than 7.8 mmol / L within any 24-hour period is 120 minutes or more; The glycosylation correction risk score is mapped to the disease risk level.

[0010] As a preferred embodiment of the diabetes risk assessment method described in this invention, the control assessment model calculates the basic control risk level based on the lifespan correction model and the glycosylation accumulation model. The control risk level is obtained by correcting the basic control risk level through control condition audit, including obtaining the corrected glycated hemoglobin data according to the lifespan correction model and obtaining the glycosylation accumulation amount according to the glycosylation accumulation model. Based on the diagnostic criteria for diabetes, thresholds for glycated hemoglobin (HbA1c) data and cumulative glycosylation (CG) were set. The corrected HbA1c data was mapped to the long-term control level according to the HbA1c data threshold. The cumulative GG was mapped to the cumulative exposure level according to the cumulative GG threshold. Both the long-term control level and the cumulative exposure level were divided into three levels: low, medium, and high. The higher level between the long-term control level and the cumulative exposure level was taken as the basic control risk level. The basic control risk level was revised through control condition audit: When the absolute value of the saccharification deviation is greater than or equal to 0.6%, the basic control risk level is adjusted according to the positive or negative direction of the saccharification deviation. If the saccharification deviation is positive, the basic control risk level is increased by one level to obtain the control risk level; if the saccharification deviation is negative, the basic control risk level remains unchanged and the control risk level is equal to the basic control risk level. If the absolute value of the saccharification deviation is less than 0.6%, no correction will be made, and the risk level will be equal to the basic control risk level.

[0011] A diabetes risk assessment system employing any of the methods described in this invention, comprising: a data acquisition module that collects user data, obtains diabetes diagnostic criteria, and performs population determination based on the user data and diabetes diagnostic criteria; The calculation module establishes lifetime correction models, glycosylation deviation models, and glycosylation accumulation models to process user data and obtain analytical data. In the disease incidence assessment module, when the population is determined to be non-diabetic, a disease incidence assessment model is established. The disease incidence assessment model calculates the basic risk score based on the lifespan correction model. The basic risk score is corrected through disease incidence condition audit to obtain the glycosylation-corrected risk score. The glycosylation-corrected risk score is then mapped to the disease incidence risk level. The control assessment module establishes a control assessment model when a population is identified as having diabetes. The control assessment model calculates the basic control risk level based on the lifespan correction model and the glycosylation accumulation model. The basic control risk level is then corrected through control condition audits to obtain the final control risk level.

[0012] The beneficial effects of this invention are as follows: This invention proposes a method for diabetes risk assessment, constructing a multi-model fusion assessment system composed of a lifespan correction model, a glycation deviation model, and a glycation accumulation model. This achieves unified modeling for the risk assessment of first-onset diabetes in non-diabetic individuals and the risk assessment of diabetes control in diabetic individuals. The system automatically distinguishes user types through a population determination module and calls either the disease assessment channel or the control assessment channel to perform risk calculations based on the determination results. This invention can simultaneously output disease risk levels and control risk levels, forming a continuous assessment chain covering the three stages of health, latent disease, and disease. This method fully utilizes the combined characteristics of blood glucose time-series data and glycated hemoglobin data, overcoming the limitations of single-indicator assessment and significantly improving the accuracy and individual adaptability of risk assessment. It also possesses scalability and can be widely applied in home portable testing environments or medical institution monitoring systems, possessing significant practical application value for early warning of diabetes, disease tracking, and personalized health management. Attached Figure Description

[0013] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 The above is an overall flowchart of a diabetes risk assessment method provided in Embodiment 1 of the present invention. Detailed Implementation

[0015] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0016] Example 1, referring to Figure 1 As an embodiment of the present invention, a method for assessing diabetes risk is provided, comprising: S1: Collect user data, obtain diabetes diagnostic criteria, and perform population determination based on user data and diabetes diagnostic criteria.

[0017] Furthermore, user data is collected and diagnostic criteria for diabetes are obtained. The user data includes glycated hemoglobin data, fasting blood glucose data, and blood glucose time-series data. The diagnostic criteria for diabetes are numerical thresholds for diagnosing diabetes. These diagnostic criteria are internationally recognized medical diagnostic standards, and the data originates from the World Health Organization. These are commonly used numerical thresholds for diagnosing diabetes in this field and are directly applied in this invention.

[0018] Specifically, glycated hemoglobin (HbA1c) data is obtained from a single test. The test can be performed at a medical facility or using a registered HbA1c testing device. Record the test results and date, in percentage form.

[0019] Fasting blood glucose data are obtained after fasting for at least 8 hours prior to blood collection. The test can be performed using the venous blood enzyme method or by finger-prick blood glucose measurement using a portable blood glucose meter. Record the test value and date, in mmol / L.

[0020] Blood glucose time-series data is used for subsequent model calculations in this invention and is not used as the basis for population determination in this step. This data is collected in one of the following two ways, and either method can be selected during implementation: Continuous glucose meter testing method: Install a continuous glucose monitoring device, set a fixed sampling interval, and continuously record blood glucose values ​​from the time of installation. Each record includes the date, time, and blood glucose value in mmol / L, forming blood glucose time series data.

[0021] Point-based testing method: Use a portable blood glucose meter to measure blood glucose at fixed times every day. Take one measurement at each of the four times: before breakfast, before lunch, before dinner, and before bedtime. Record the date, time, and blood glucose value for each measurement in mmol / L. Sort the measurement results by time to form a blood glucose time series data.

[0022] Furthermore, by comparing user data with diabetes diagnostic criteria, population segmentation is completed: When fasting blood glucose levels are ≥7.0 mmol / L or glycated hemoglobin levels are ≥6.5%, the population is classified as having diabetes. When fasting blood glucose is <7.0 mmol / L and glycated hemoglobin is <6.5%, the population is considered non-diabetic.

[0023] All of the above portable devices can be used to perform tests independently in a home environment, without relying on the laboratory conditions of medical institutions.

[0024] S2: Establish lifetime correction models, glycosylation deviation models, and glycosylation accumulation models to process user data and obtain analytical data.

[0025] The average blood glucose level is obtained by calculating the arithmetic mean of the blood glucose values ​​in the blood glucose time series data.

[0026] Lifetime correction models are used to eliminate the influence of erythrocyte lifespan differences on glycated hemoglobin (HbA1c) data. Specifically, HbA1c formation is determined by both blood glucose levels and erythrocyte lifespan. For each individual, erythrocyte lifespan remains relatively constant and can be considered a fixed value. Therefore, under a given average blood glucose level, the rate of HbA1c formation is a fixed value.

[0027] First, calculate the glycation rate coefficient based on the average blood glucose level: ; in, Indicates the saccharification rate coefficient; , Represents the linear coefficients of saccharification kinetics, constants; The average blood glucose level is represented by the glycation rate coefficient, which characterizes the amount of glycation produced per unit time at that average blood glucose level.

[0028] Then, by correlating the glycated hemoglobin data with the glycation rate coefficient, the equivalent red blood cell lifespan matching the glycated hemoglobin data is obtained: ; Where L represents the equivalent red blood cell lifespan; This represents glycated hemoglobin data.

[0029] Finally, the glycated hemoglobin data was standardized and converted using the baseline red blood cell lifespan to obtain the corrected glycated hemoglobin data: ; in, This indicates the corrected glycated hemoglobin data; This represents the baseline red blood cell lifespan, which is 120 days. Corrected glycated hemoglobin data ensures that subsequent calculations are performed under uniform lifespan conditions, avoiding systematic bias caused by individual differences in red blood cell lifespan.

[0030] It should be noted that the baseline red blood cell lifespan refers to the average physiological lifespan of red blood cells from generation to clearance in healthy, non-anemic adults. In this invention, the baseline red blood cell lifespan is used to standardize the glycated hemoglobin (HbA1c) results of different individuals, that is, to convert the measured HbA1c value to the HbA1c level assuming a normal red blood cell lifespan, thereby eliminating systematic bias caused by individual lifespan differences. The average red blood cell lifespan of healthy adults is approximately 100–120 days; in this embodiment, the baseline red blood cell lifespan is set to 120 days.

[0031] The glycated hemoglobin (HbA1c) deviation model is used to quantify whether HbA1c data deviates from its expected level given a mean blood glucose level. A stable one-to-one correspondence exists between mean blood glucose and HbA1c data, allowing the model to derive the expected HbA1c data through regression. Specifically, the mean blood glucose unit is converted from mmol / L to mg / dL, and the model first derives the expected HbA1c data based on the mean blood glucose level. ; in, This indicates the expected glycated hemoglobin data; The average blood glucose level is represented by the unit conversion; then, the measured glycated hemoglobin data is subtracted from the expected glycated hemoglobin data to obtain the glycated deviation value.

[0032] Glycation deviation (GD) is a quantitative result: a positive value indicates higher Glycation levels at the same average blood glucose level, while a negative value indicates lower Glycation levels at the same average blood glucose level. GD is used to characterize individual differences in Glycation tendency: in morbidity assessment models, it serves as an enhancement to morbidity condition auditing to identify individuals at higher risk at the same average blood glucose level; in control assessment models, it serves as a trigger for control condition auditing to identify cases where individuals have long met control targets but exhibit abnormal Glycation tendency.

[0033] The glycosylation accumulation model takes blood glucose time series data and glycated hemoglobin data as inputs. Based on the diagnostic criteria for diabetes, the hyperglycemia threshold is set at 7.8 mmol / L (the general numerical threshold for postprandial hyperglycemia). The super-threshold exposure integral is obtained using a discrete and feasible calculation process, and then linearly synthesized with the glycation level to form a unified quantity that can be directly used for assessment.

[0034] Arrange the blood glucose time series data in chronological order, and use the blood glucose value in the i-th blood glucose time series data as... This indicates that the timestamp is Protein glycation is a concentration-dependent, non-enzymatic reaction. Blood glucose levels above the physiological safety limit will drive the formation of glycation products at a higher rate. Therefore, only the portion exceeding the threshold is cumulatively calculated, which can avoid equating fluctuations within the normal range with damage contributions.

[0035] Specifically, using 7.8 mmol / L as the threshold, the non-negative portion exceeding the threshold is taken. The suprathreshold amplitudes of adjacent points are weighted using a trapezoidal method and multiplied by the corresponding time difference to obtain the suprathreshold exposure integral. ; Where E represents the over-threshold exposure integral; N represents the total number of blood glucose time series data; i represents the i-th data point in the blood glucose time series data; max represents taking the maximum value; This represents the blood glucose value in the i-th blood glucose time series data; This represents the blood glucose value in the (i+1)th blood glucose time series data; express and The time difference.

[0036] The long-term accumulation of damage depends on the product of the non-negative portion exceeding the threshold and time. Therefore, a time-weighted integral is performed on the non-negative portion exceeding the threshold to obtain E, which is then normalized over the entire duration to obtain the average overthreshold exposure intensity, used to characterize the average intensity of sustained hyperglycemia during the statistical period. Linear synthesis is performed on the average overthreshold exposure intensity and glycated hemoglobin data to obtain the cumulative glycation amount: ; Where A represents the cumulative amount of glycosylation; α represents the contribution weight of the glycated hemoglobin portion; H represents the glycated hemoglobin data; and β represents the contribution weight of the hyperglycemic exposure portion. This represents the average intensity of exposure beyond the threshold.

[0037] It should be noted that, to ensure the dimensional consistency and reasonable contribution ratio of glycated hemoglobin data and average over-threshold exposure intensity, the coefficients α and β were calibrated offline during the model establishment phase. Specifically, a user sample dataset with complete blood glucose time-series data and glycated hemoglobin test records was collected. The average over-threshold exposure intensity and glycated hemoglobin data for each sample were obtained, along with the clinically detected levels of advanced glycation end products (corresponding to cumulative glycation). With the objective of minimizing the mean squared error between the model output cumulative glycation (calculated using the method of this invention) and the level of advanced glycation end products, linear regression was used to obtain the values ​​of coefficients α and β. The obtained α and β are fixed constants that remain unchanged during model operation.

[0038] Glycated hemoglobin data reflect the average glycation level on a 2–3 month timescale, while the average super-threshold exposure intensity reflects the average intensity of high exposure during the statistical period. Both data characterize the accumulation mechanism from the perspectives of background glycation and peak exposure, respectively. A single cumulative amount of glycation is formed through the linear synthesis of α and β, enabling the deterministic expression of long-term background and peak exposure in the same index.

[0039] The superthreshold exposure integral quantitatively represents the cumulative level of above-threshold blood glucose during the statistical period; the mean superthreshold exposure intensity provides a characterization of high exposure intensity aligned with unit time, which can distinguish the situation of "not high average but frequent peaks"; glycosylation accumulation combines long-term glycation level with high exposure intensity into a single input, which is used in control assessment models to reflect the risk associated with cumulative damage, and in morbidity assessment models to identify individual differences with high exposure characteristics when average levels are close.

[0040] User data is processed using a lifetime correction model, a glycation deviation model, and a glycation accumulation model to output analytical data, which includes corrected glycated hemoglobin data, glycation deviation values, and cumulative glycation amounts.

[0041] S3: When the population is determined to be non-diabetic, an morbidity assessment model is established. The morbidity assessment model calculates the basic risk score based on the lifespan correction model. The basic risk score is corrected through morbidity condition audit to obtain the glycosylation-corrected risk score. The glycosylation-corrected risk score is then mapped to the morbidity risk level.

[0042] Furthermore, the corrected glycated hemoglobin data are obtained based on the lifespan correction model, and the glycated deviation value is obtained based on the glycated deviation model.

[0043] Using the corrected glycated hemoglobin data as the basic input to the morbidity assessment model, the Sigmoid function is used to map the corrected glycated hemoglobin data to a range of 0 to 1, thereby obtaining the baseline risk score: ; The basic risk score represents the probability of a risk occurring, with a value range of [0,1]. To avoid boundary effects and nonlinear imbalances in probability space calculations, it is converted into a probability form: ; in, This indicates the basic risk score; This represents the Sigmoid function; k represents the control slope. This represents the corrected glycated hemoglobin data; c represents the center point of the risk curve; This represents the probability corresponding to the base risk score. Probability represents the relative ratio of an event occurring to its non-occurrence, facilitating multiplicative or exponential adjustments.

[0044] It should be noted that the center point of the risk curve is set at 6.5% based on the internationally accepted diagnostic criteria for diabetes. Referring to large-scale epidemiological cohort studies (such as UKPDS and DCCT), the relationship between glycated hemoglobin data and the risk of diabetic complications is non-linear, ranging from 6.0% to 8.0%. Therefore, based on the user sample dataset, the control slope is obtained by minimizing the difference between the baseline risk score and the actual incidence rate, so that the corrected glycated hemoglobin data is consistent with the upward trend of the actual incidence risk rate.

[0045] The baseline risk score was adjusted by auditing the conditions for the onset of the disease. According to long-term epidemiological follow-up studies (UKPDS, DCCT, Hempe, etc.), when the absolute value of the glycated deviation is less than 0.5%, the deviation is mainly affected by test fluctuations and short-term physiological factors. When the absolute value of the glycated deviation is greater than or equal to 0.6%, it indicates that the individual's glycated tendency is significantly higher than the population average, which has independent physiological significance and is significantly correlated with an increased risk of diabetic complications. Therefore, this invention uses an absolute value of 0.6% for the glycated deviation as the trigger condition for auditing the pathogenesis. If the absolute value of the glycated deviation is greater than or equal to 0.6%, the glycated deviation is retrieved, the baseline risk score is corrected, and the corrected glycated risk score is output. If the absolute value of the glycated deviation is less than 0.6%, no correction is performed, and the corrected glycated risk score is equal to the baseline risk score.

[0046] Glycated hemoglobin (HbA1c) deviation refers to the degree to which an individual's HbA1c data is higher or lower than the expected HbA1c data under a given average blood glucose level. In logistic regression, the HbA1c deviation has a linear effect on the log-odds ratio, which is naturally consistent with risk interpretation. Specifically, the probability corresponding to the base risk score is multiplicatively corrected using an exponential function, and the risk amplification coefficient is calculated using the deviation value as input, thus achieving dynamic adjustment of the risk score. The correction formula is: ; in, This indicates the probability corresponding to the glycation correction risk score; This indicates the probability corresponding to the basic risk score; Represents an exponential function; The saccharification correction risk score gain coefficient is represented by P, which represents the saccharification deviation value. The saccharification correction risk score is determined by... The reverse calculation yields: ; in, This indicates the glycation correction risk score.

[0047] The average amplification factor of the risk of developing diabetes was determined by regression analysis of glycation deviation values ​​and the incidence of diabetes in a user sample dataset.

[0048] By applying an exponential function to the probability form of the base risk score, the impact of saccharification deviation on risk is made to change monotonically and continuously. When the saccharification deviation is positive, the exponential term is greater than 1, and the risk score is adjusted upward; when the saccharification deviation is negative, the exponential term is less than 1, and the risk score remains the same or decreases slightly, thus achieving automated dynamic correction based on the degree of individual saccharification deviation.

[0049] The glycation correction risk score is adjusted based on the cumulative amount of glycosylation. When any glycosylation correction condition is met, the cumulative amount of glycosylation is compressed using the Sigmoid function to obtain an intensity score. A gating coefficient is calculated based on the glycation correction risk score. The glycosylation correction risk score is calculated and output based on the gating coefficient and the intensity score. If the glycosylation correction condition is not met, no correction is performed, and the glycosylation correction risk score is equal to the glycation correction risk score.

[0050] The glycosylation correction conditions include two: Blood glucose time series data with a continuous duration of greater than 7.8 mmol / L for 30 minutes or more; The cumulative duration of blood glucose time series data greater than 7.8 mmol / L within any 24-hour period is 120 minutes or more.

[0051] During the correction process, the accumulated glycosylation is first compressed using a sigmoid function, converting it into an intensity score within the 0-1 range to limit the excessive amplification of the risk score by extremely high values. Then, a gating coefficient is calculated based on the glycosylation-corrected risk score; this coefficient equals "1 minus the glycosylation-corrected risk score" and reflects the potential for upward adjustment of the risk. Finally, combining the gating coefficient and the intensity score, a multiplicative correction is performed using a preset gain coefficient to calculate and output the glycosylation-corrected risk score. The correction formula is as follows: ; in, This indicates the probability corresponding to the glycosylation correction risk score; This represents the gain coefficient of the glycosylation correction risk score; The Sigmoid compression result represents the cumulative amount of glycosylation. This indicates the glycation correction risk score.

[0052] Convert the probability corresponding to the glycosylation correction risk score into a glycosylation correction risk score: ; in, This indicates the risk score for glycosylation correction.

[0053] The 7.8 mmol / L threshold in the glycation correction conditions is derived from the international standards for postprandial hyperglycemia set by the World Health Organization (WHO) and the American Diabetes Association (ADA); the time thresholds of 30 consecutive minutes and 120 minutes cumulatively over 24 hours are derived from international blood glucose monitoring guidelines, reflecting the duration limits at which sustained hyperglycemic exposure may accelerate glycation. These two conditions ensure that the model can accurately capture high-exposure events that have a real impact on long-term glycation accumulation.

[0054] The cumulative amount of glycation reflects both the long-term average glycation level and the intensity of high glycemic exposure, exhibiting a dual cumulative effect on glycation reactions in terms of both time and magnitude. By using sigmoid compression to suppress extreme values ​​and by dynamically adjusting the correction magnitude through a gating coefficient, the enhancement of high exposure is made more significant in low-baseline-risk individuals, while remaining stable in high-risk individuals. This ensures the stability of the assessment while achieving a sensitive response of risk assessment to individuals with high exposure.

[0055] The glycosylation-corrected risk score is mapped to a disease risk level, and intervals are divided according to preset thresholds. Specifically, the glycosylation-corrected risk score intervals are divided into three levels: low risk, medium risk, and high risk. For example, a glycosylation-corrected risk score below 0.4 is considered low risk, between 0.4 and 0.7 is considered medium risk, and above 0.7 is considered high risk. Finally, based on the glycosylation-corrected risk score interval, the corresponding disease risk level is output.

[0056] S4: When a population is identified as having diabetes, a control assessment model is established. The control assessment model is based on the lifespan correction model and the glycosylation accumulation model to calculate the basic control risk level. The basic control risk level is then corrected through control condition audits to obtain the control risk level.

[0057] Furthermore, the control assessment model is based on a lifespan-corrected model and a cumulative glycosylation model. Glycated hemoglobin (HbA1c) data and cumulative glycosylation levels jointly reflect the long-term control level of diabetic patients and the cumulative effect of hyperglycemia exposure. The lifespan-corrected model eliminates measurement bias caused by differences in red blood cell lifespan, ensuring comparability of HbA1c data across individuals. The cumulative glycosylation model integrates blood glucose time-series data with glycosylation levels, reflecting the cumulative effect of the duration and magnitude of hyperglycemia exposure, thus characterizing the quality of long-term glycemic control. By combining these two models, the baseline control risk level considers both long-term average glycosylation levels and glycemic fluctuations and exposure intensity.

[0058] Specifically, the corrected glycated hemoglobin data are obtained based on the lifespan correction model, and the cumulative amount of glycation is obtained based on the cumulative glycation model.

[0059] Based on the diagnostic criteria for diabetes, thresholds for glycated hemoglobin (HbA1c) data and cumulative glycosylation (CG) were set. The corrected HbA1c data was mapped to the long-term control level according to the HbA1c data threshold. The cumulative GG was mapped to the cumulative exposure level according to the cumulative GG threshold. Both the long-term control level and the cumulative exposure level were divided into three levels: low, medium, and high. The higher level between the long-term control level and the cumulative exposure level was taken as the basic control risk level.

[0060] The glycated hemoglobin (HbA1c) thresholds were set based on the World Health Organization (WHO) and the American Diabetes Association (ADA) diabetes control standards. 7.0% was the primary target for glycemic control; below 7.0% indicated good control with a low-risk long-term control level; 7.0%–8.0% indicated moderate control with a medium-risk long-term control level; and above 8.0% indicated poor control with a high-risk long-term control level. The cumulative glycation level threshold was determined statistically using a user sample dataset. Based on the fitted relationship between cumulative glycation level and complication risk rate, the inflection point where the complication risk rate significantly increases was used as the grading boundary, resulting in a three-tiered classification of low, medium, and high risk.

[0061] The basic control risk level was revised through control condition audit: When the absolute value of the saccharification deviation is greater than or equal to 0.6%, the basic control risk level is adjusted according to the positive or negative direction of the saccharification deviation. If the saccharification deviation is positive, the basic control risk level is increased by one level to obtain the control risk level; if the saccharification deviation is negative, the basic control risk level remains unchanged and the control risk level is equal to the basic control risk level.

[0062] If the absolute value of the saccharification deviation is less than 0.6%, no correction will be made, and the control risk level will be equal to the basic control risk level.

[0063] The control condition audit correction mechanism is used to dynamically adjust the control risk level when glycated hemoglobin (HbA1c) deviation exists. When the absolute value of the HbA1c deviation is greater than or equal to 0.6%, it indicates a difference beyond the normal fluctuation range between the measured HbA1c level and the mean blood glucose level. A positive HbA1c deviation value indicates a higher tendency for HbA1c at the same mean blood glucose level, suggesting a more active long-term HbA1c process and increased potential control risk; therefore, the baseline control risk level is raised by one level. A negative HbA1c deviation value indicates that the individual's HbA1c level is slightly lower than expected, which falls within the scope of physiological differences and requires no adjustment.

[0064] This invention constructs a risk assessment system based on glycated hemoglobin (HbA1c) data and blood glucose time-series data, achieving unified modeling for both first-onset diabetes risk assessment in non-diabetic individuals and blood glucose control risk assessment in diabetic individuals within the same method. The method automatically identifies population types based on users' blood glucose and HbA1c characteristics, and outputs disease risk and control risk levels respectively, enabling simultaneous assessment of health risks at different stages. This provides a unified, objective, and quantifiable assessment method for early warning and later management of diabetes.

[0065] Example 2, in an exemplary embodiment, also provides a diabetes risk assessment system, including, The data collection module collects user data, obtains diabetes diagnostic criteria, and performs population determination based on user data and diabetes diagnostic criteria.

[0066] The calculation module establishes a lifetime correction model, a glycosylation deviation model, and a glycosylation accumulation model based on user data.

[0067] In the disease incidence assessment module, when the population is determined to be non-diabetic, a disease incidence assessment model is established. The disease incidence assessment model calculates a basic risk score based on the life expectancy correction model. The basic risk score is corrected through disease incidence condition audit to obtain a glycosylation-corrected risk score, which is then mapped to a disease incidence risk level.

[0068] The control assessment module establishes a control assessment model when a population is identified as having diabetes. The control assessment model calculates the basic control risk level based on the lifespan correction model and the glycosylation accumulation model. The basic control risk level is then corrected through control condition audits to obtain the final control risk level.

[0069] If the above functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0070] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0071] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0072] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0073] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for assessing the risk of diabetes, characterized in that, include: Collect user data, obtain diabetes diagnostic criteria, and perform population determination based on user data and diabetes diagnostic criteria; Establish lifetime correction models, glycosylation deviation models, and glycosylation accumulation models; process user data, and obtain analytical data. When the population is determined to be non-diabetic, an incidence assessment model is established. The incidence assessment model calculates the basic risk score based on the lifespan correction model. The basic risk score is corrected by the audit of the incidence conditions to obtain the glycosylation-corrected risk score. The glycosylation-corrected risk score is then mapped to the incidence risk level. When a population is identified as having diabetes, a control assessment model is established. The control assessment model calculates the basic control risk level based on the lifespan correction model and the glycosylation accumulation model. The basic control risk level is then corrected through a control condition audit to obtain the control risk level.

2. The diabetes risk assessment method as described in claim 1, characterized in that: The process of collecting user data, obtaining diabetes diagnostic criteria, and performing population determination based on user data and diabetes diagnostic criteria includes the following: the user data includes glycated hemoglobin data, fasting blood glucose data, and blood glucose time series data. The stated diagnostic criteria for diabetes are numerical thresholds for diagnosing diabetes. User data is compared with these diagnostic criteria to determine the target population. When fasting blood glucose levels are ≥7.0 mmol / L or glycated hemoglobin levels are ≥6.5%, the population is classified as having diabetes. When fasting blood glucose is <7.0 mmol / L and glycated hemoglobin is <6.5%, the population is considered non-diabetic.

3. The diabetes risk assessment method as described in claim 2, characterized in that: The lifespan correction model includes: calculating the glycation rate coefficient based on average blood glucose, calculating the equivalent red blood cell lifespan based on glycated hemoglobin data and the glycation rate coefficient, correcting the glycated hemoglobin data based on the equivalent red blood cell lifespan, and outputting the corrected glycated hemoglobin data. Establishing a glycated hemoglobin deviation model involves calculating the expected glycated hemoglobin data based on the average blood glucose level, subtracting the glycated hemoglobin data from the glycated hemoglobin data, and outputting the glycated deviation value. The establishment of the glycosylation accumulation model includes setting the hyperglycemia threshold to 7.8 mmol / L based on the diagnostic criteria for diabetes, and calculating the excess exposure integral based on blood glucose time series data: ； Where E represents the over-threshold exposure integral; N represents the total number of blood glucose time series data; i represents the i-th data point in the blood glucose time series data; max represents taking the maximum value; This represents the blood glucose value in the i-th blood glucose time series data; This represents the blood glucose value in the (i+1)th blood glucose time series data; express and Time difference; The over-threshold exposure integral is normalized over the entire duration to calculate the average over-threshold exposure intensity. The average over-threshold exposure intensity and glycated hemoglobin data are linearly synthesized to obtain the cumulative amount of glycosylation. The data analyzed included corrected glycated hemoglobin data, glycation deviation values, and cumulative glycation.

4. The diabetes risk assessment method as described in claim 3, characterized in that: The disease incidence assessment model calculates a basic risk score based on a lifespan-corrected model. This basic risk score is then corrected through an audit of disease incidence conditions to obtain a glycosylation-corrected risk score. This glycosylation-corrected risk score is then mapped to disease incidence risk levels. The corrected glycated hemoglobin data are obtained based on the lifespan correction model, and the glycated deviation value is obtained based on the glycated deviation model. Using the corrected glycated hemoglobin data as the basic input of the disease assessment model, the Sigmoid function is used to map the corrected glycated hemoglobin data to a range of 0 to 1 to obtain the basic risk score; The baseline risk score was adjusted by auditing the conditions for the onset of the disease. The basic risk score is adjusted based on the saccharification deviation value. If the absolute value of the saccharification deviation value is greater than or equal to 0.6%, the saccharification deviation value is called, and the basic risk score is adjusted by an exponential function to output the saccharification-corrected risk score. If the absolute value of the saccharification deviation value is less than 0.6%, no adjustment is made, and the saccharification-corrected risk score is equal to the basic risk score. The glycation correction risk score is adjusted based on the cumulative amount of glycosylation. When any glycosylation correction condition is met, the cumulative amount of glycosylation is compressed using the Sigmoid function to obtain an intensity score. The glycation correction risk score is calculated based on the glycation correction risk score. The glycation correction risk score is calculated and output based on the glycation correction coefficient and the intensity score. If the glycosylation correction condition is not met, no correction is performed, and the glycation correction risk score is equal to the glycation correction risk score. The glycosylation correction conditions include two: Blood glucose time series data with a continuous duration of greater than 7.8 mmol / L for 30 minutes or more; The cumulative duration of blood glucose time series data greater than 7.8 mmol / L within any 24-hour period is 120 minutes or more; The glycosylation correction risk score is mapped to the disease risk level.

5. The diabetes risk assessment method as described in claim 4, characterized in that: The control assessment model calculates the basic control risk level based on the lifetime correction model and the glycosylation accumulation model. The control risk level is obtained by correcting the basic control risk level through control condition audit. This includes obtaining the corrected glycated hemoglobin data based on the lifetime correction model and obtaining the glycosylation accumulation amount based on the glycosylation accumulation model. Based on the diagnostic criteria for diabetes, thresholds for glycated hemoglobin (HbA1c) data and cumulative glycosylation (CG) were set. The corrected HbA1c data was mapped to the long-term control level according to the HbA1c data threshold. The cumulative GG was mapped to the cumulative exposure level according to the cumulative GG threshold. Both the long-term control level and the cumulative exposure level were divided into three levels: low, medium, and high. The higher level between the long-term control level and the cumulative exposure level was taken as the basic control risk level. The basic control risk level was revised through control condition audit: When the absolute value of the saccharification deviation is greater than or equal to 0.6%, the basic control risk level is adjusted according to the positive or negative direction of the saccharification deviation. If the saccharification deviation is positive, the basic control risk level is increased by one level to obtain the control risk level. If the saccharification deviation value is negative, the basic control risk level remains unchanged, and the control risk level is equal to the basic control risk level. If the absolute value of the saccharification deviation is less than 0.6%, no correction will be made, and the risk level will be equal to the basic control risk level.

6. A diabetes risk assessment system, applied to the diabetes risk assessment method according to any one of claims 1 to 5, characterized in that, include, The data collection module collects user data, obtains diabetes diagnostic criteria, and performs population determination based on user data and diabetes diagnostic criteria. The calculation module establishes lifetime correction models, glycosylation deviation models, and glycosylation accumulation models to process user data and obtain analytical data. In the disease incidence assessment module, when the population is determined to be non-diabetic, a disease incidence assessment model is established. The disease incidence assessment model calculates the basic risk score based on the lifespan correction model. The basic risk score is corrected through disease incidence condition audit to obtain the glycosylation-corrected risk score. The glycosylation-corrected risk score is then mapped to the disease incidence risk level. The control assessment module establishes a control assessment model when a population is identified as having diabetes. The control assessment model calculates the basic control risk level based on the lifespan correction model and the glycosylation accumulation model. The basic control risk level is then corrected through control condition audits to obtain the final control risk level.

Images