Method and apparatus for predicting data trend of blood glucose monitoring device
By constructing a linear DTS error grid and trend consistency matrix, the shortcomings of blood glucose monitoring devices in assessing the accuracy of single-point values and trends are addressed, enabling horizontal comparison between devices and risk avoidance, thereby improving the comprehensiveness and accuracy of blood glucose monitoring.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-05-07
- Publication Date
- 2026-07-14
AI Technical Summary
Existing blood glucose monitoring devices cannot evaluate the accuracy of blood glucose values based on trends, and it is difficult to make horizontal comparisons between different devices. Furthermore, the failure to incorporate risk rules based on differences in blood glucose ranges leads to insufficient identification or redundant warnings.
By constructing a DTS error grid with linear boundaries and combining it with continuous reference blood glucose data, the single-point accuracy and trend accuracy of discrete blood glucose monitoring data are evaluated, and a trend consistency matrix is established to achieve a comprehensive accuracy evaluation of blood glucose monitoring devices.
It enables quantitative evaluation of the single-point accuracy and trend accuracy of blood glucose monitoring devices, enhancing the comprehensiveness and accuracy of device evaluation, adapting to actual application scenarios with different blood glucose concentration ranges, and avoiding problems of insufficient risk identification and early warning.
Smart Images
Figure CN122392967A_ABST
Abstract
Description
Technical Field
[0001] This application belongs to the fields of medical technology and computer technology, and specifically relates to a method and device for predicting the data change trend of a blood glucose monitoring device. Background Technology
[0002] With the rapid development of blood glucose monitoring technology, fingertip blood glucose monitors and continuous glucose monitoring devices have been widely used in the management and clinical monitoring of corresponding diseases. Current methods for assessing the accuracy of blood glucose monitoring typically employ error grid analysis, which compares device measurements with reference values and evaluates the reliability of the monitoring data based on the magnitude of the difference. This method only assesses the accuracy of single-point blood glucose values and does not establish a standardized quantitative correlation between numerical accuracy indicators and clinical risk indicators, resulting in the inability to make cross-sectional comparisons of assessment results between different devices and studies.
[0003] Continuous glucose monitoring (CGM) devices can be used to assist insulin administration and blood glucose management. However, they lack differentiated risk rules for high and low blood glucose ranges, leading to insufficient risk identification in low-glucose ranges and redundant warnings in high-glucose ranges. In summary, current technologies struggle to simultaneously meet the integrated assessment requirements of numerical accuracy and trend accuracy, exhibiting significant deficiencies in applicability and accuracy. Summary of the Invention
[0004] This application provides a method and device for predicting the data change trend of a blood glucose monitoring device. The purpose is to solve the problem that existing blood glucose monitoring assessments can only evaluate single-point values and cannot assess the accuracy of trends. At the same time, it solves the problem that it is difficult to make horizontal comparisons between different devices, and the problem of insufficient identification and early warning caused by not judging risks according to blood glucose interval differences.
[0005] In a first aspect, embodiments of this application provide a method for predicting the data change trend of a blood glucose monitoring device, the method comprising: Acquire discrete blood glucose data output by the blood glucose monitoring device, as well as synchronously collected continuous reference blood glucose data; Based on the discrete blood glucose monitoring data and continuous reference blood glucose data, a linear boundary DTS error grid is constructed, and the data proportion of the target area is determined according to the distribution of data in the DTS error grid. The average absolute relative difference is determined based on the data proportion of the target area, and the single-point accuracy assessment of discrete blood glucose monitoring data is completed based on the average absolute relative difference. Reference trend data is calculated based on the time-series changes of the continuous reference blood glucose data, and monitoring trend indicators are determined based on the time-series changes of the discrete monitoring blood glucose data. The reference trend data and the monitoring trend indicators are mapped to preset trend levels respectively, and a trend consistency matrix is constructed based on the trend level correspondence. The trend accuracy of the blood glucose monitoring device is evaluated based on the degree of consistency between the monitored trend indicators and the reference trend data in the trend consistency matrix, and the evaluation results are output.
[0006] Furthermore, based on the discrete blood glucose monitoring data and the continuous reference blood glucose data, a linear boundary DTS error grid is constructed, including: The criteria for risk zoning are determined based on the numerical deviation between discrete blood glucose monitoring data and continuous reference blood glucose data. Based on the risk zoning criteria, a nonlinear risk boundary for the initial error grid is generated; The nonlinear risk boundary is smoothly fitted using a continuous function, transforming it into a linear boundary. The coordinate system is divided into risk areas based on the linear boundary to form the DTS error grid.
[0007] Furthermore, the step of using a continuous function to smoothly fit the nonlinear risk boundary includes: Curve fitting of the nonlinear risk boundary of the initial error grid is performed using a continuous function; Based on the fitting results, the jagged distribution characteristics of the boundary are eliminated, and the nonlinear boundary is normalized into a straight segmented boundary.
[0008] Furthermore, determining the data proportion of the target area based on the distribution of data in the DTS error grid includes: The discrete blood glucose monitoring data and the continuous reference blood glucose data are combined to form a data pair and mapped to the DTS error grid; Count the number of data pairs that fall into the target area and the total number of data pairs; The percentage of data in the target area is determined by the ratio of the number of data pairs falling into the target area to the total number of data pairs.
[0009] Furthermore, determining the average absolute relative difference based on the data proportion of the target area includes: A single-point accuracy quantification model is established based on a pre-defined linear transformation relationship; The data proportion of the target area is input into the quantification model to calculate the average absolute relative difference.
[0010] Furthermore, the calculation of reference trend data based on the time-series changes of continuous reference blood glucose data includes: Obtain the current continuous reference blood glucose value and the preset historical continuous reference blood glucose values; The magnitude of blood glucose trend change is determined based on the difference between the current reference blood glucose value and the preset historical reference blood glucose value. The corresponding reference trend data is determined based on the magnitude of the trend change.
[0011] Furthermore, the mapping of reference trend data and monitoring trend indicators to preset trend levels includes: Based on the direction and rate of blood glucose changes, establish a multi-level trend level classification rule; According to the trend level division rules, the reference trend data and the monitoring trend indicators are mapped to the corresponding trend levels respectively.
[0012] Furthermore, the construction of a trend consistency matrix based on the trend level correspondence includes: Construct a matrix coordinate system with the reference trend level as the horizontal axis and the monitoring trend level as the vertical axis; Match each trend level combination to the corresponding cell in the matrix; The frequency of trend combinations within each cell is counted to form the trend consistency matrix.
[0013] Furthermore, the assessment of the trend accuracy of the blood glucose monitoring device includes: Based on the trend consistency matrix, the frequency of combinations with consistent trend levels is statistically analyzed. A quantitative assessment of trend accuracy is generated based on the ratio of the frequency of combinations with consistent trend levels to the total frequency of combinations.
[0014] Secondly, embodiments of this application provide a data change trend prediction device for a blood glucose monitoring device, the device comprising: The blood glucose data acquisition module is used to acquire discrete blood glucose data output by the blood glucose monitoring device, as well as synchronously collected continuous reference blood glucose data. The data proportion determination module is used to construct a linear boundary DTS error grid based on the discrete blood glucose monitoring data and the continuous reference blood glucose data, and to determine the data proportion of the target area based on the distribution of data in the DTS error grid. The single-point assessment module is used to determine the average absolute relative difference based on the data proportion of the target area, and to complete the single-point accuracy assessment of discrete blood glucose monitoring data based on the average absolute relative difference. The trend calculation module is used to calculate reference trend data based on the time-series changes of the continuous reference blood glucose data, and to determine the monitoring trend index based on the time-series change amplitude of the discrete monitoring blood glucose data. The consistency matrix construction module is used to map the reference trend data and the monitoring trend indicators to preset trend levels respectively, and construct a trend consistency matrix according to the trend level correspondence. The evaluation result output module is used to evaluate the trend accuracy of the blood glucose monitoring device based on the degree of consistency between the monitoring trend indicators and the reference trend data in the trend consistency matrix, and output the evaluation result.
[0015] Thirdly, embodiments of this application provide an electronic device including a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein the program or instructions, when executed by the processor, implement the steps of the method described in the first aspect.
[0016] Fourthly, embodiments of this application provide a readable storage medium on which a program or instructions are stored, which, when executed by a processor, implement the steps of the method described in the first aspect.
[0017] Fifthly, embodiments of this application provide a chip, the chip including a processor and a communication interface, the communication interface being coupled to the processor, the processor being used to run programs or instructions to implement the method as described in the first aspect.
[0018] The technical solution provided in this application constructs a linear boundary DTS error grid to quantitatively assess the accuracy of discrete blood glucose data at single points, establishes a correlation between clinical risk indicators and numerical evaluation indicators, and effectively ensures the horizontal comparability of evaluation results from different blood glucose monitoring devices. Simultaneously, it calculates reference trend data and monitoring trend indicators by combining time-series data, and builds a trend consistency matrix based on standardized trend levels to objectively determine the consistency of trend changes in blood glucose monitoring devices. Furthermore, it completes a comprehensive assessment of trend accuracy through the degree of trend matching, overcoming the deficiency of existing technologies that cannot simultaneously assess single-point values and changing trends. Moreover, this technical solution can be adapted to practical application scenarios with different blood glucose concentration ranges, avoiding insufficient risk identification and early warning issues, and significantly improving the comprehensiveness and accuracy of the overall evaluation of blood glucose monitoring devices. Attached Figure Description
[0019] Figure 1 This is a flowchart illustrating the method for predicting data change trends in a blood glucose monitoring device provided in Embodiment 1 of this application; Figure 2 This is a schematic diagram of the distribution of the first blood glucose meter provided in this application in the DTS error grid; Figure 3 This is a schematic diagram of the distribution of the second blood glucose meter provided in this application within the DTS error grid; Figure 4 This is a schematic diagram of the distribution of the continuous glucose monitor provided in this application in the DTS error grid; Figure 5 This is a schematic diagram of the trend accuracy matrix structure provided in this application; Figure 6 This is a schematic diagram of the data change trend prediction device for the blood glucose monitoring equipment provided in Embodiment 2 of this application; Figure 7This is a schematic diagram of the structure of the electronic device provided in Embodiment 3 of this application. Detailed Implementation
[0020] To make the objectives, technical solutions, and advantages of this application clearer, specific embodiments of this application will be described in further detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely for explaining this application and not for limiting it. It should also be noted that, for ease of description, only the parts relevant to this application are shown in the drawings, not all of them. Before discussing exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe operations (or steps) as sequential processes, many of these operations can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but may also have additional steps not included in the drawings. The process can correspond to a method, function, procedure, subroutine, subroutine, etc.
[0021] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.
[0022] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class and the number of objects is not limited; for example, a first object can be one or more. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.
[0023] The following description, in conjunction with the accompanying drawings, details the data change trend prediction method, apparatus, equipment, and medium for blood glucose monitoring devices provided in this application, through specific embodiments and application scenarios.
[0024] Example 1 Figure 1 This is a flowchart illustrating the method for predicting data change trends in a blood glucose monitoring device provided in Embodiment 1 of this application. Figure 1 As shown, the specific steps include the following: S11, acquire discrete blood glucose monitoring data output by the blood glucose monitoring device, as well as synchronously acquired continuous reference blood glucose data.
[0025] Discrete blood glucose monitoring data can be fragmented blood glucose test values collected and output at intervals by blood glucose monitoring devices, single-point blood glucose data obtained by fingertip blood glucose meters at regular intervals, or periodic blood glucose data captured by continuous monitoring devices at fixed time intervals.
[0026] Continuous reference blood glucose data can be standard blood glucose values collected continuously throughout the process and used as a benchmark for accuracy comparison, or benchmark data collected at high frequency by high-precision medical testing instruments, or continuous monitoring standard blood glucose data after clinical calibration.
[0027] In this solution, two types of blood glucose data are retrieved and received via a terminal. Specifically, data transmission can be completed via Bluetooth, wireless network, or by directly reading the collected data synchronously stored in the local cache.
[0028] S12, Based on the discrete blood glucose monitoring data and the continuous reference blood glucose data, construct a DTS error grid with a linear boundary, and determine the data proportion of the target area based on the distribution of data in the DTS error grid.
[0029] The DTS error grid (DTS, Dynamic Trend Assessment System) refers to a two-dimensional assessment grid used to compare the deviation between monitored values and reference values and to divide risk intervals. It is a rectangular coordinate grid that corresponds to blood glucose values on two axes and is a dedicated error grid adapted to customized zoning of blood glucose ranges.
[0030] The target area refers to the pre-defined low-error, low-risk judgment area within the DTS error grid. For example, it can be a Class A safety area used to characterize the qualified test results, or a limited area within the allowable error range defined according to clinical standards.
[0031] This solution can define grid boundaries and partitions based on two sets of blood glucose data, determine partition limits according to the range of numerical deviation, and also complete grid area division by combining clinical risk level.
[0032] S13, determine the average absolute relative difference based on the data proportion of the target area, and complete the single-point accuracy assessment of discrete blood glucose monitoring data based on the average absolute relative difference.
[0033] The average absolute relative difference is an evaluation parameter that quantifies the overall deviation of a single group of blood glucose test data. It can be a relative error index that reflects the deviation of the monitored value from the standard value.
[0034] Single-point accuracy assessment refers to the determination of the reasonableness of error in a single independent blood glucose test result, which can be the classification of the deviation level of discrete test values.
[0035] This solution can be used to perform calculations according to fixed conversion rules, or it can automatically match and obtain the corresponding parameters through a preset quantization model.
[0036] S14, calculate reference trend data based on the time-series changes of the continuous reference blood glucose data, and determine monitoring trend indicators based on the time-series change amplitude of the discrete monitoring blood glucose data.
[0037] Reference trend data refers to standardized trend data generated based on the continuous baseline blood glucose time-series fluctuation pattern. It can be continuous change curve data of blood glucose rise and fall in a short period of time, or statistical data of blood glucose time-series fluctuation obtained through a certain standardization method.
[0038] Monitoring trend indicators refer to graded indicators that characterize the direction and fluctuation range of blood glucose changes detected by the device. They can be basic trend indicators that distinguish between rising, stable, and falling trends, or multi-level trend labels used to classify the speed of change.
[0039] Specifically, the fluctuation status can be determined by calculating the numerical difference between two adjacent moments, or by combining the overall fluctuation amplitude within a time period.
[0040] S15, map the reference trend data and the monitoring trend indicators to preset trend levels respectively, and construct a trend consistency matrix based on the trend level correspondence.
[0041] Preset trend levels are standardized levels that uniformly classify blood sugar change characteristics. They can be multi-level adjustment levels divided according to the rate of change, such as three basic levels that can be set to simplify the needs of trend comparison.
[0042] A trend consistency matrix is a two-dimensional statistical table used to statistically compare the reference trend with the monitoring trend. It can be a frequency statistical matrix corresponding to two-dimensional levels, such as a trend comparison matrix used for deviation classification statistics.
[0043] Specifically, the corresponding level can be matched according to the change characteristics, or the gear level can be uniformly converted by using a threshold.
[0044] S16. Based on the degree of consistency between the monitored trend indicators and the reference trend data in the trend consistency matrix, evaluate the trend accuracy of the blood glucose monitoring device and output the evaluation result.
[0045] Consistency refers to the degree of matching between the monitored trend and the standard reference trend in terms of the direction of change and the rhythm of fluctuation. It can be the proportion of synchronous overlap between the two sets of trends, or the degree of consistency of trend characteristics.
[0046] Trend accuracy refers to the ability of a blood glucose monitoring device to accurately capture the dynamic changes in blood glucose levels. This can be the device's ability to reproduce trends over time or the overall adaptability of dynamic blood glucose monitoring.
[0047] In this scheme, the proportion of matched samples can be statistically analyzed for quantitative evaluation, and the overall performance can be determined by combining the deviation distribution.
[0048] This technical solution achieves quantitative evaluation of single-point data by simultaneously acquiring two types of blood glucose data and combining them with a linear boundary DTS error grid. It also introduces time-series trend analysis and matrix comparison modes to simultaneously complete the dual detection and evaluation of numerical accuracy and trend changes. This overcomes the shortcomings of existing technologies that can only evaluate single-point data and cannot consider trend determination. It establishes a unified quantitative evaluation standard, facilitating horizontal comparisons across multiple devices and effectively improving the comprehensiveness and clinical applicability of blood glucose monitoring and evaluation.
[0049] In one embodiment, optionally, a linear boundary DTS error grid is constructed based on the discrete blood glucose monitoring data and the continuous reference blood glucose data, including: The criteria for risk zoning are determined based on the numerical deviation between discrete blood glucose monitoring data and continuous reference blood glucose data. Based on the risk zoning criteria, a nonlinear risk boundary for the initial error grid is generated; The nonlinear risk boundary is smoothly fitted using a continuous function, transforming it into a linear boundary. The coordinate system is divided into risk areas based on the linear boundary to form the DTS error grid.
[0050] Numerical deviation refers to the difference between the monitored blood glucose value and the reference blood glucose value at the same time period. It can be the absolute value of positive or negative error, or the relative deviation value after normalization.
[0051] The criteria for risk zoning refer to the evaluation standards for classifying different risk levels based on the error range. These standards can be clinically prescribed error thresholds or the grading limits corresponding to the equipment's detection accuracy.
[0052] Nonlinear risk boundaries can be irregularly shaped, curved partition boundaries in traditional error grids, such as tortuous boundaries formed by fitting the original data, or irregularly trimmed natural partition edges.
[0053] This solution can correct the boundary shape through function fitting. Specifically, it can perform segmented regularization of the curve boundary or uniformly correct it into a standardized straight line segmented structure.
[0054] This technical solution establishes a partitioning basis based on data deviation, optimizes and rectifyes traditional nonlinear and irregular boundaries, and uniformly transforms curved boundaries into straight-line partition boundaries. This effectively solves the problems of jagged and chaotic original grid boundaries and partitioning, making the risk area division more regular and uniform, and improving the accuracy of subsequent data classification and statistics.
[0055] In one embodiment, optionally, the smooth fitting of the nonlinear risk boundary using a continuous function includes: Curve fitting of the nonlinear risk boundary of the initial error grid is performed using a continuous function; Based on the fitting results, the jagged distribution characteristics of the boundary are eliminated, and the nonlinear boundary is normalized into a straight segmented boundary.
[0056] Continuous functions are operational functions that can achieve curve fitting and edge correction. They can be polynomial fitting functions or continuous processing functions such as smoothing filters.
[0057] The jagged distribution feature can be the edge shape of the original grid boundary, which is interspersed with concave and convex features, discontinuous and messy. For example, it can be the irregular undulating edge caused by the discrete distribution of data, or the fine and tortuous boundary generated by the original fitting.
[0058] A linear segmented boundary is a partition boundary formed by a combination of multiple regular straight lines, and can be a segmented standardized straight line boundary.
[0059] This solution can smooth out edge defects through fitting operations, such as through overall smoothing and noise reduction, or segmented local correction and optimization.
[0060] This technical solution defines the specific implementation method for boundary smoothing fitting, specifically eliminates the irregular defects of the original boundary, unifies the shape of the partition boundary, reduces data misclassification and misjudgment caused by boundary clutter, and enhances the overall standardization of the DTS error grid and its adaptability to clinical scenarios.
[0061] In one embodiment, optionally, determining the data proportion of the target area based on the distribution of data in the DTS error grid includes: The discrete blood glucose monitoring data and the continuous reference blood glucose data are combined to form a data pair and mapped to the DTS error grid; Count the number of data pairs that fall into the target area and the total number of data pairs; The percentage of data in the target area is determined by the ratio of the number of data pairs falling into the target area to the total number of data pairs.
[0062] A data pair refers to a set of monitoring data and reference data that match each other at the same time point. It can be time-synchronized dual-source blood glucose paired data or corresponding comparison data after period alignment.
[0063] Paired data can be labeled to grid coordinate positions, such as by matching coordinates according to double values, or by batch importing data to complete a unified coordinate mapping, and then the proportion of regional data can be obtained through statistical calculations.
[0064] This technical solution presents a calculation process for the target area proportion. It completes grid mapping and quantity statistics based on synchronized pairing data. The calculation logic is clear and rigorous, and the statistical results are accurate and reliable, providing stable data support for subsequent deviation parameter calculation and single-point accuracy evaluation.
[0065] In one embodiment, optionally, determining the average absolute relative difference based on the data proportion of the target area includes: A single-point accuracy quantification model is established based on a pre-defined linear transformation relationship; The data proportion of the target area is input into the quantification model to calculate the average absolute relative difference.
[0066] Among them, the linear transformation relationship refers to the fixed corresponding operation rule between the regional proportion and the deviation parameter. It can be a linear proportional conversion relationship or a piecewise linear calibration conversion rule.
[0067] A single-point accuracy quantification model refers to a computational model equipped with conversion rules for automatically calculating deviation indicators. In this solution, it can be a numerical calculation model with a built-in fixed algorithm or a calibrable intelligent quantitative evaluation model.
[0068] This technical solution establishes a standardized quantitative method to accurately correlate the proportion of low-risk areas with error evaluation parameters, bridging the gap between clinical risk indicators and numerical accuracy indicators, and facilitating comparative analysis of different models and types of blood glucose monitoring devices.
[0069] In one embodiment, optionally, calculating the reference trend data based on the time-series changes of continuous reference blood glucose data includes: Obtain the current continuous reference blood glucose value and the preset historical continuous reference blood glucose values; The magnitude of blood glucose trend change is determined based on the difference between the current reference blood glucose value and the preset historical reference blood glucose value. The corresponding reference trend data is determined based on the magnitude of the trend change.
[0070] The current reference blood glucose value is the latest set of baseline blood glucose test values collected in real time. It can be the latest baseline data sampled immediately, or it can be the end-time series data that is updated in real time.
[0071] Preset historical time points are pre-defined past comparison time points, such as historical time points twenty minutes before the current time, or custom retrospective time points adapted to blood sugar change cycles.
[0072] The magnitude of trend change is the difference in blood glucose values between adjacent time points. It can be a small fluctuation in the short term or an overall change over a long period of time.
[0073] This technical solution objectively derives the standard blood glucose change trend by analyzing the difference between baseline data before and after the time series, ensuring the uniformity and objectivity of the reference trend data, and providing a standard and reliable reference for subsequent equipment monitoring trend comparison.
[0074] In one embodiment, optionally, mapping the reference trend data and the monitoring trend indicator to a preset trend level includes: Based on the direction and rate of blood glucose changes, establish a multi-level trend level classification rule; According to the trend level division rules, the reference trend data and the monitoring trend indicators are mapped to the corresponding trend levels respectively.
[0075] Among them, the direction of change refers to the overall trend of blood glucose fluctuations, which can be a continuous rise in blood glucose, or a decrease or stable trend in blood glucose.
[0076] The rate of change refers to how quickly blood glucose levels increase or decrease per unit of time. It can be a high rate of change with rapid fluctuations or a low rate of change with slow fluctuations.
[0077] Multi-level trend level classification rules can be standardized grading standards formulated in combination with fluctuation characteristics, such as comprehensive grading rules that take into account both direction and rate, or other basic level classification rules that simplify classification.
[0078] This technical solution clarifies the grading standards for two types of trend data, transforms differentiated fluctuation data into standardized grading information, reduces data dimensional differences, lowers the computational difficulty of trend consistency comparison, and improves the overall efficiency of trend analysis.
[0079] In one embodiment, optionally, constructing a trend consistency matrix based on the trend level correspondence includes: Construct a matrix coordinate system with the reference trend level as the horizontal axis and the monitoring trend level as the vertical axis; Match each trend level combination to the corresponding cell in the matrix; The frequency of trend combinations within each cell is counted to form the trend consistency matrix.
[0080] Among them, the matrix coordinate system refers to a two-dimensional statistical framework composed of the horizontal axis and the vertical axis. It can be an orthogonal coordinate system with two corresponding positions, or other customized statistical coordinate frameworks.
[0081] Trend gear combination refers to the pairing combination of reference gear and monitoring gear within the same detection period. It can be a randomly matched dual-trend combination or a one-to-one corresponding combination with time synchronization.
[0082] Frequency of occurrence is the statistical number of samples in which the same trend combination occurs repeatedly. It can be the cumulative frequency of samples tested in a single day or the statistical frequency of multiple periods.
[0083] This technical solution provides a complete and specific method for constructing a trend consistency matrix. Based on a two-dimensional level, it realizes trend combination classification and statistics, which can intuitively present the trend matching and deviation distribution, and provide an intuitive and statistical data carrier for the quantitative assessment of trend accuracy.
[0084] In one embodiment, optionally, the assessment of the trend accuracy of the blood glucose monitoring device includes: Based on the trend consistency matrix, the frequency of combinations with consistent trend levels is statistically analyzed. A quantitative assessment of trend accuracy is generated based on the ratio of the frequency of combinations with consistent trend levels to the total frequency of combinations.
[0085] Among them, the trend level consistent combination refers to the paired sample where the reference trend level and the monitoring trend level are completely consistent. It can be a trend combination where the direction of change and the rate of change are completely matched, or it can be a synchronous trend sample where the level division is completely unified.
[0086] Total combination frequency refers to the total number of trend-paired samples participating in the statistics. It can be the total number of samples in a single testing period or the total number of statistical samples aggregated from multiple batches.
[0087] Quantitative evaluation results refer to digital evaluation conclusions obtained through percentage calculations. These can be trend matching degrees in percentage form or trend performance evaluation results with hierarchical annotations.
[0088] This technical solution provides a novel quantitative analysis method using matrix statistical data, replacing traditional qualitative descriptions with objective numerical values to achieve accurate evaluation of trend monitoring capabilities. This facilitates intuitive assessment of the device's dynamic blood glucose capture performance and provides data support for device optimization and clinical selection.
[0089] To enable those skilled in the art to better understand this solution, this application also provides a preferred embodiment.
[0090] The technical solution proposed in this invention aims to solve the following problems existing in current blood glucose monitoring accuracy assessment tools: Existing error grids (such as Clark Error Grid CEG, Parkes Error Grid PEG, and Monitoring Error Grid SEG) are only applicable to blood glucose monitoring systems (BGMs) and cannot reflect the clinical accuracy of continuous blood glucose monitoring systems (CGMs), which have become an important tool for blood glucose monitoring.
[0091] The Monitoring Error Grid (SEG) has a usability problem. It defines 5, 9, or 15 risk regions with non-linear boundaries, which means that nearby data pairs representing the same relative error may be assigned to different risk regions. This does not conform to clinical practice and is difficult to use.
[0092] There is a lack of tools for assessing the accuracy of CGM trends. Existing tools either do not involve trend assessment or are too complex to be widely adopted, such as Continuous Glucose Error Grid (CG-EGA) analysis. However, trend information of CGMs (the rate and direction of blood glucose changes) is crucial for clinical decisions such as insulin dosage adjustment in diabetic patients.
[0093] The lack of a clear relationship between clinical accuracy (as measured by the percentage of the DTS error grid in area A) and analytical accuracy (as measured by the mean absolute relative difference MARD) makes it difficult to directly compare results from different studies.
[0094] This technical solution achieves a comprehensive and accurate evaluation of blood glucose monitoring devices (BGMs and CGMs) by optimizing existing error grids, unifying evaluation standards, establishing trend evaluation tools, and clarifying indicator correlations. The specific working process is as follows: Step 1: Optimize SEG to create DTS error grid: Based on the consensus of 89 international experts, the risk area boundaries of SEG are smoothed, changing nonlinear boundaries to linear boundaries, forming a DTS error grid containing five risk areas (A: no risk; B: slight risk; C: moderate risk; D: high risk; E: extreme risk), making it suitable for point accuracy assessment of both BGMs and CGMs. The rationality of the area adjustment is verified by analyzing 5542 data pairs from 18 BGMs (only 2.6% of the data involved switching between the A and B intervals of the SEG and DTS error grids).
[0095] Step 2: Establish the relationship between the percentage of data in area A and MARD: Analyze the accuracy study data of 22 BGM and 9 CGM studies, derive and verify the mathematical relationship between the percentage of data in area A (pZA) and the mean absolute relative difference (MARD) in the DTS error grid, and obtain the formula:
[0096] This enables a direct conversion between clinical accuracy and analytical accuracy metrics. MARD, short for Mean Absolute Relative Difference, is a crucial indicator for assessing the analytical accuracy of blood glucose monitoring devices. pZA, short for percentage of reference / monitor pairs in Zone A, represents the percentage of reference / monitor data pairs located in Zone A of the DTS error grid. pZA is an important indicator for measuring the clinical accuracy of blood glucose monitoring devices. A pZA of 90% roughly corresponds to a 10% mean absolute relative deviation (MARD). For every 1% increase in pZA, MARD decreases by 0.33%; conversely, for every 1% decrease in MARD, pZA increases.
[0097] Step 3: Create the DTS trend accuracy matrix: Define the CGM trend index (+2 to -2, representing different rates of blood glucose change) and the reference trend calculation method (based on the difference between the current value and the reference value 15-45 minutes ago), construct a 5×5 trend consistency matrix, and divide the trend error into 5 risk categories (Category 1: no risk; Category 2: slight underestimation risk; Category 3: moderate overestimation risk; Category 4: severe underestimation risk; Category 5: extreme overestimation risk), and adjust the risk classification criteria according to the reference blood glucose values (<100mg / dL, 100-180mg / dL, >180mg / dL).
[0098] Figure 2 This is a schematic diagram of the distribution of the first blood glucose meter provided in this application within the DTS error grid. Figure 3 This is a schematic diagram of the distribution of the second blood glucose meter provided in this application within the DTS error grid. Figure 4 This is a schematic diagram of the distribution of the continuous glucose monitor provided in this application within the DTS error grid. For example... Figure 2 , Figure 3 as well as Figure 4 As shown, with the reference blood glucose value on the horizontal axis and the monitored blood glucose value on the vertical axis, five risk zones (AE) are divided using straight line boundaries. The coordinate axes support conversion between mg / dL and mmol / L units. For example, 89.7% of the data from the first blood glucose meter is located in zone A, 99.1% of the data from the second blood glucose meter is located in zone A, and 88.3% of the data from the CGM combined dataset is located in zone A, intuitively reflecting the differences in accuracy between different devices.
[0099] The overall accuracy of the 3952 reference / monitoring datasets for continuous glucose monitoring (CGM) devices is as follows: In Zone A, the percentage of the first blood glucose meter is 89.7%, while that of the second blood glucose meter is 99.1%. The percentage of CGM devices in Zone A is 88.3%. BGM represents blood glucose meters; CGM represents continuous glucose monitoring devices.
[0100] Figure 5 This is a schematic diagram of the trend accuracy matrix structure provided in this application, such as... Figure 5 As shown, a 5×5 grid is formed with the reference trend on the horizontal axis and the CGM monitoring trend on the vertical axis. Each cell represents the frequency and risk category of a specific trend combination. For example, in the matrix for reference blood glucose of 1-600 mg / dL, 72.68% of the data belong to category 1 (no risk), and there are no category 4 or 5 risk data; in the matrix for reference blood glucose <100 mg / dL, 0.96% of the data belong to category 5 (extreme overestimation risk), reflecting the high-risk characteristic of overestimation trend when hypoglycemia occurs.
[0101] This design incorporates the following key improvements: Risk-based boundary smoothing techniques refer to fitting the SEG risk score with a continuous function to eliminate the "jagged" edges of the region boundaries, ensuring that data with the same relative error are assigned to the same risk region.
[0102] Trend risk classification rules: In line with the FDA's requirements for integrated CGM (iCGM), extreme risks (classes 4 and 5) are associated with clinical critical errors (such as CGM showing an increase in blood glucose when it is actually decreasing rapidly), thereby improving the clinical relevance of trend assessment.
[0103] Multi-scenario adaptation technology: Adjust the trend risk classification for different reference blood glucose ranges. For example, there are no five types of risks when blood glucose is high (>180mg / dL), which reflects the scenario specificity of risk assessment.
[0104] The advantages of this technical solution compared to existing technologies are as follows: Combining the structural features of the DTS error grid and trend accuracy matrix, it has the following significant advantages compared to existing technologies (CEG, PEG, SEG, etc.): 1. Wider range of applications and stronger compatibility; Existing technologies (CEG, PEG, SEG) are designed only for blood glucose monitors (BGMs) and cannot be adapted to the clinical accuracy assessment (including point accuracy and trend accuracy) of continuous glucose monitors (CGMs).
[0105] The DTS error grid of this technical solution, through the unified design of 5 risk areas (AE) and straight line boundaries, can be applied to the point accuracy assessment of both BGMs and CGMs, solving the problem of "excessive specificity and insufficient compatibility" of existing tools.
[0106] 2. Usability is significantly improved; The existing SEG has problems such as overly fine risk area division (5, 9 or 15 types of areas) and non-linear boundaries, which leads to chaotic data classification (data with the same relative error may be classified into different areas), making it difficult to use.
[0107] This technical solution uses continuous function fitting to fit the risk score, optimizing the nonlinear boundary of SEG into a straight boundary, eliminating "sawtooth" interference, and retaining only 5 risk regions, simplifying the analysis process and making it easier for clinicians, researchers, and regulatory agencies to quickly interpret the results.
[0108] 3. For the first time, a standardized assessment of trend accuracy has been achieved; Existing technologies do not address the accuracy assessment of CGM trends. The CG-EGA proposed in 2004 has not been widely adopted due to its complex analysis, resulting in a lack of unified assessment tools for CGM trend indicators (such as the rate and direction of blood glucose change).
[0109] This technical solution's DTS trend accuracy matrix uses a 5×5 matrix structure and 5 risk categories (1-5) to intuitively quantify the deviation between the CGM trend indicator and the reference trend. It also combines FDA requirements to define extreme risk scenarios (such as CGM showing an increase in blood sugar but actually a rapid decrease), filling a gap in this field.
[0110] 4. Establish a quantitative correlation between clinical and analytical accuracy; In existing technologies, there is no clear correlation between clinical accuracy (such as error grid region distribution) and analytical accuracy (such as MARD), making it difficult to compare results from different studies.
[0111] This technical solution analyzes 22 BGM and 9 CGM studies to derive the mathematical relationship between pZA (percentage in zone A) and MARD (for every 1% increase in pZA, MARD decreases by approximately 0.33%), achieving direct conversion between the two indicators and providing a unified standard for cross-study comparisons.
[0112] 5. Risk assessment is more aligned with clinical practice; The risk classification of existing technologies does not take into account scenario specificity (such as higher error risk during hypoglycemia).
[0113] The trend accuracy matrix of this technical solution adjusts the risk classification for different reference blood glucose ranges (<100mg / dL, 100-180mg / dL, >180mg / dL): it expands the judgment range of extreme overestimation risk (5 categories) in the case of hypoglycemia, and cancels the 5 categories of risk in the case of hyperglycemia, so that the assessment is more in line with the needs of clinical decision-making.
[0114] 6. The tools are publicly available and easy to use; Some existing error grid tools have limitations or high complexity.
[0115] The DTS error grid and trend accuracy matrix of this technical solution are in the public domain. The accompanying browser software supports unit conversion between mg / dL and mmol / L, allows customization of chart parameters (such as axis range and point size), and can export results (PNG or PDF format), making it convenient for developers, researchers, and regulatory agencies to use.
[0116] In summary, this technical solution, through structural optimization, functional expansion, and standardized design, outperforms existing technologies in terms of compatibility, ease of use, and clinical relevance, providing a more reliable evaluation tool for the research, development, certification, and clinical application of blood glucose monitoring devices.
[0117] Example 2 Figure 6 This is a schematic diagram of the data change trend prediction device for the blood glucose monitoring equipment provided in Embodiment 2 of this application. Figure 6 As shown, the device includes: The blood glucose data acquisition module 601 is used to acquire discrete blood glucose data output by the blood glucose monitoring device, as well as synchronously acquired continuous reference blood glucose data. The data proportion determination module 602 is used to construct a linear boundary DTS error grid based on the discrete blood glucose monitoring data and the continuous reference blood glucose data, and determine the data proportion of the target area based on the distribution of data in the DTS error grid. The single-point assessment module 603 is used to determine the average absolute relative difference based on the data proportion of the target area, and to complete the single-point accuracy assessment of discrete blood glucose monitoring data based on the average absolute relative difference. The trend calculation module 604 is used to calculate reference trend data based on the time-series changes of the continuous reference blood glucose data, and to determine the monitoring trend index based on the time-series change amplitude of the discrete monitoring blood glucose data. The consistency matrix construction module 605 is used to map the reference trend data and the monitoring trend indicators to preset trend levels respectively, and construct a trend consistency matrix according to the trend level correspondence. The evaluation result output module 606 is used to evaluate the trend accuracy of the blood glucose monitoring device based on the degree of consistency between the monitoring trend indicators and the reference trend data in the trend consistency matrix, and output the evaluation result.
[0118] The data change trend prediction device for blood glucose monitoring equipment in this application embodiment can be a system, or a component, integrated circuit, or chip in a terminal. The system can be a mobile electronic device or a non-mobile electronic device. For example, mobile electronic devices can be mobile phones, tablets, laptops, PDAs, in-vehicle electronic devices, wearable devices, ultra-mobile personal computers (UMPCs), netbooks, or personal digital assistants (PDAs), etc., while non-mobile electronic devices can be servers, network attached storage (NAS), personal computers (PCs), televisions (TVs), ATMs, or self-service machines, etc. This application embodiment does not impose specific limitations.
[0119] The data change trend prediction device for blood glucose monitoring equipment in this embodiment can be a device with an operating system. This operating system can be Android, iOS, or other possible operating systems; this embodiment does not specifically limit the specific operating system.
[0120] The data change trend prediction device for blood glucose monitoring equipment provided in this application embodiment can realize the various processes of the above embodiments, and will not be described again here to avoid repetition.
[0121] Example 3 like Figure 7 As shown, this application embodiment also provides an electronic device 700, including a processor 701, a memory 702, and a program or instructions stored in the memory 702 and executable on the processor 701. When the program or instructions are executed by the processor 701, they implement the various processes of the above-described blood glucose monitoring device data change trend prediction method embodiment and achieve the same technical effect. To avoid repetition, they will not be described again here.
[0122] It should be noted that the electronic devices in the embodiments of this application include mobile electronic devices and non-mobile electronic devices as described above.
[0123] Example 4 This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described method embodiment for predicting the data change trend of a blood glucose monitoring device and achieve the same technical effect. To avoid repetition, they will not be described again here.
[0124] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.
[0125] Example 5 This application also provides a chip, which includes a processor and a communication interface. The communication interface is coupled to the processor. The processor is used to run programs or instructions to implement the various processes of the above-described method embodiment for predicting data change trends of blood glucose monitoring devices, and can achieve the same technical effect. To avoid repetition, it will not be described again here.
[0126] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.
[0127] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element. Furthermore, it should be noted that the scope of the methods and systems in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.
[0128] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a computer software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application.
[0129] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.
[0130] The above description is merely a preferred embodiment and the technical principles employed in this application. This application is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions that can be made by those skilled in the art will not depart from the scope of protection of this application. Therefore, although this application has been described in detail through the above embodiments, this application is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of this application, the scope of which is determined by the scope of the claims.
Claims
1. A method for predicting the data change trend of a blood glucose monitoring device, characterized in that, The method includes: Acquire discrete blood glucose data output by the blood glucose monitoring device, as well as synchronously collected continuous reference blood glucose data; Based on the discrete blood glucose monitoring data and continuous reference blood glucose data, a linear boundary DTS error grid is constructed, and the data proportion of the target area is determined according to the distribution of data in the DTS error grid. The average absolute relative difference is determined based on the data proportion of the target area, and the single-point accuracy assessment of discrete blood glucose monitoring data is completed based on the average absolute relative difference. Reference trend data is calculated based on the time-series changes of the continuous reference blood glucose data, and monitoring trend indicators are determined based on the time-series changes of the discrete monitoring blood glucose data. The reference trend data and the monitoring trend indicators are mapped to preset trend levels respectively, and a trend consistency matrix is constructed based on the trend level correspondence. The trend accuracy of the blood glucose monitoring device is evaluated based on the degree of consistency between the monitored trend indicators and the reference trend data in the trend consistency matrix, and the evaluation results are output.
2. The method according to claim 1, characterized in that, Based on the discrete blood glucose monitoring data and continuous reference blood glucose data, a linear boundary DTS error grid is constructed, including: The criteria for risk zoning are determined based on the numerical deviation between discrete blood glucose monitoring data and continuous reference blood glucose data. Based on the risk zoning criteria, a nonlinear risk boundary for the initial error grid is generated; The nonlinear risk boundary is smoothly fitted using a continuous function, transforming it into a linear boundary. The coordinate system is divided into risk areas based on the linear boundary to form the DTS error grid.
3. The method according to claim 2, characterized in that, The method of using a continuous function to smoothly fit the nonlinear risk boundary includes: Curve fitting of the nonlinear risk boundary of the initial error grid is performed using a continuous function; Based on the fitting results, the jagged distribution characteristics of the boundary are eliminated, and the nonlinear boundary is normalized into a straight segmented boundary.
4. The method according to claim 2, characterized in that, Determining the data proportion of the target area based on the distribution of data in the DTS error grid includes: The discrete blood glucose monitoring data and the continuous reference blood glucose data are combined to form a data pair and mapped to the DTS error grid; Count the number of data pairs that fall into the target area and the total number of data pairs; The percentage of data in the target area is determined by the ratio of the number of data pairs falling into the target area to the total number of data pairs.
5. The method according to claim 1, characterized in that, The determination of the average absolute relative difference based on the data proportion of the target area includes: A single-point accuracy quantification model is established based on a pre-defined linear transformation relationship; The data proportion of the target area is input into the quantification model to calculate the average absolute relative difference.
6. The method according to claim 1, characterized in that, The calculation of reference trend data based on the time-series changes of continuous reference blood glucose data includes: Obtain the current continuous reference blood glucose value and the preset historical continuous reference blood glucose values; The magnitude of blood glucose trend change is determined based on the difference between the current reference blood glucose value and the preset historical reference blood glucose value. The corresponding reference trend data is determined based on the magnitude of the trend change.
7. The method according to claim 1, characterized in that, The step of mapping reference trend data and monitoring trend indicators to preset trend levels includes: Based on the direction and rate of blood glucose changes, establish a multi-level trend level classification rule; According to the trend level division rules, the reference trend data and the monitoring trend indicators are mapped to the corresponding trend levels respectively.
8. The method according to claim 7, characterized in that, The construction of a trend consistency matrix based on the trend level correspondence includes: Construct a matrix coordinate system with the reference trend level as the horizontal axis and the monitoring trend level as the vertical axis; Match each trend level combination to the corresponding cell in the matrix; The frequency of trend combinations within each cell is counted to form the trend consistency matrix.
9. The method according to claim 1, characterized in that, The assessment of the trend accuracy of the blood glucose monitoring device includes: Based on the trend consistency matrix, the frequency of combinations with consistent trend levels is statistically analyzed. A quantitative assessment of trend accuracy is generated based on the ratio of the frequency of combinations with consistent trend levels to the total frequency of combinations.
10. A data change trend prediction device for a blood glucose monitoring device, characterized in that, The device includes: The blood glucose data acquisition module is used to acquire discrete blood glucose data output by the blood glucose monitoring device, as well as synchronously collected continuous reference blood glucose data. The data proportion determination module is used to construct a linear boundary DTS error grid based on the discrete blood glucose monitoring data and continuous reference blood glucose data, and determine the data proportion of the target area based on the distribution of data in the DTS error grid. The single-point assessment module is used to determine the average absolute relative difference based on the data proportion of the target area, and to complete the single-point accuracy assessment of discrete blood glucose monitoring data based on the average absolute relative difference. The trend calculation module is used to calculate reference trend data based on the time-series changes of the continuous reference blood glucose data, and to determine the monitoring trend index based on the time-series change amplitude of the discrete monitoring blood glucose data. The consistency matrix construction module is used to map the reference trend data and the monitoring trend indicators to preset trend levels respectively, and construct a trend consistency matrix according to the trend level correspondence. The evaluation result output module is used to evaluate the trend accuracy of the blood glucose monitoring device based on the degree of consistency between the monitoring trend indicators and the reference trend data in the trend consistency matrix, and output the evaluation result.