System and method for identifying clinically similar clusters of continuous glucose monitoring (CGM) profiles

Abstract

Embodiments relate to a system for processing glucose data with efficient glucose database management. The system includes a physical data store that includes glucose measurement data and a representation for at least one cluster of glucose measurement data, the representation approximating a blood glucose profile vector array for one cluster of a plurality of glucose profiles segmented by a plurality of time ranges. The system includes a processor and a computer memory configured by stored instructions, the instructions, when executed, causing the processor to: 1) receive glucose measurement values; 2) convert the glucose measurement values into vector form; 3) search the physical data store by comparing a newly received glucose measurement value with the centroid of a cluster using a similarity measure; 3) classify the newly received glucose measurement value by a cluster having a matching similarity measure based on the comparison; and 4) attribute a treatment to the newly received glucose measurement value.

Description

Technical Field

[0001] (Cross - Reference to Related Applications) This patent application is a related application to U.S. Provisional Application No. 63 / 448,918, filed on February 7, 2023, and U.S. Provisional Application No. 63 / 335,361, filed on April 27, 2022, and claims the benefit of their priorities. The entire contents of each of these provisional applications are incorporated herein by reference.

[0002] (Technical Field) Embodiments relate to systems for processing glucose data through efficient glucose database management and using the classified glucose data to monitor, analyze, and affect the concentration of glucose levels in body fluids.

Background Art

[0003] Glucose variability (GV) in diabetes reflects a biological behavioral process that causes fluctuations in blood glucose (BG), having two main dimensions: namely, amplitude representing the degree of BG excursions and time representing the frequency of BG fluctuations and the rate of event progression. Over the past 20 years, the ability to observe this process has evolved from occasional self - monitoring (e.g., several BG determinations per day) to the latest continuous glucose monitoring (CGM) that incorporates high - density datasets of BG readings at equally spaced intervals in time (e.g., every 5 minutes). These datasets, well - known as time series, open up new possibilities for the analysis and optimal control of the human metabolic system in diabetes, including the evaluation of system dynamics, prediction of BG trends and events such as impending hypoglycemia or hyperglycemia, and what is commonly referred to as "artificial pancreas" or automatic closed - loop control.

[0004] The widespread adoption of CGM technology inevitably generates large amounts of data. For example, the inventors' latest report on the real-life use of an artificial pancreas system was based on over a billion data points. The diabetes data ecosystem plays an increasingly important role in supporting data sharing, virtual clinics, and remote access. Cloud databases store this data and require the use of data science tools such as pattern recognition, neural networks, deep learning, and artificial intelligence, all of which can contribute to improving treatment and creating fully automated systems. The most promising application of cloud databases and data science tools is the use of adaptive technologies that can "learn" and personalize treatment for each individual. To do this, it is necessary to create a simple, finite, and fixed appropriate structure in the CGM data space that accurately represents the clinical meaning of the CGM data profile and at the same time does not require changing the structure for each new dataset.

[0005] There are multiple measures of glycemic control, which are described in detail in a 2017 paper published by the inventors in Nature Reviews Endocrinology. CGM-based measures generally must include not only the amplitude of CGM readings but also a certain concept of the timing of CGM readings. Some of the existing measures such as MAGE (mean amplitude of glycemic excursions) and LBGI / HBGI (low and high BG indices) have been similarly adapted for CGM use, and the adaptation of MAGE to CGM data follows the traditional time-dependent structure of this measure. Thus, in this case, CGM is used only as a source of amplitude assessment, and the adaptation of LBGI and HBGI is applied to the differences between SMBG and CGM data. The mean of daily differences (MODD) was introduced as a measure of daytime variability, and continuous overlapping net glycemic action (CONGA) was presented as a composite index of the magnitude and timing of BG fluctuations captured over various periods. The standard deviation of the rate of change of BG has been used as a marker of the stability of the metabolic system over time, based on the premise that abnormal BG changes are signs of system instability. An array of standard deviations was introduced to reflect the GV included within various clinically relevant periods of CGM data, and the clinical interpretation of various CGM-based measures of glucose variability was discussed. A review of the statistical methods available for the analysis of CGM data included several graphs such as the Poincaré plot of system stability and the variability rate-grid analysis (VGA) used to visualize the glycemic fluctuations captured by GCM

[12] . VGA was also used to show the effectiveness of closed-loop control algorithms [4]

[13] . The ideas published in Diabetes Care re-evaluated some of the methods of computing and visualizing GV in the context of the relationship between GV and the risk of hypoglycemia, and the inventors show the readers in this document the details of the interpretation of VGA and the Poincaré plot of CGM data.

[0006] The field of CGM is burdened with not only vast and complex datasets but also numerous metrics used to evaluate various aspects of CGM profiles. Therefore, the 2019 international consensus on time in range (TIR, generally 70 - 180 mg / dL), to which the inventors also contributed, set clinical goals to propose and use TIR as the primary CGM - based metric for glycemic control. In the past three years, the "TIR metric system" has been widely adopted. The TIR system is based on the ambulatory glucose profile (AGP), which has been introduced as a template for data presentation and visualization. Originally developed by Mazz et al., the standardized CGM report incorporates core CGM metrics and target values along with a 14 - day synthetic glucose profile as an essential element for clinical decision - making. This recommendation has been approved by the international consensus and is also referenced by the 2019 American Diabetes Association Standards of Medical Care in Diabetes and the American Association of Clinical Endocrinologists Consensus in the use of CGM. The AGP report is currently adopted by many CGM device manufacturers in CGM companion software. An example of the AGP report and the TIR metric system is shown in Figure 2.

[0007] The TIR metric system defines five time - in - range values for blood glucose levels. These time - in - range values are used in addition to the AGP to provide a numerical interpretation of the AGP plot. In one embodiment, for example, these time - in - range values are: level 2 hypoglycemia - below 54 mg / dL, level 1 hypoglycemia - 54 to 69 mg / dL, target range (TIR) - 70 - 180 mg / dL, level 1 hyperglycemia - 180 to 250 mg / dL, and level 2 hyperglycemia - above 250 mg / dL. Other embodiments of TIR following consensus recommendations for various types of diabetes are shown in Figure 3.

[0008] As can be seen from FIGS. 2 and 3, neither the AGP nor the TIR scale system represents the daytime variability of the CGM trace, nor does it provide a fixed finite structure for a number of daily CGM profiles. Aspects of embodiments of the systems, methods, and computer-readable media of the present invention proceed to the next step.

Prior Art Documents

Patent Documents

[0009]

Patent Document 1

Patent Document 2

Summary of the Invention

Means for Solving the Problems

[0010] Embodiments can relate to a system for processing glucose data by efficient glucose database management. The system can include a physical data store that includes glucose measurement data and a representation for at least one cluster of glucose measurement data. The representation can approximate a blood glucose profile vector for one cluster of a plurality of glucose profiles segmented by a plurality of time ranges. The system can include a processor and a computer memory configured to store instructions that, when executed, cause the processor to perform any of the method steps disclosed herein. The instructions can cause the processor to receive glucose measurement values. The instructions can cause the processor to convert the glucose measurement values into vector form. The instructions can cause the processor to search the physical data store by comparing newly received glucose measurement values to the centroid of the cluster using a similarity metric. The instructions can cause the processor to classify newly received glucose measurement values by a cluster having a matching similarity metric based on the comparison. The instructions can cause the processor to attribute a treatment to the newly received glucose measurement values.

[0011] Embodiments can relate to a method of processing glucose data for efficient glucose database management. The method can include receiving glucose measurements. The method can include converting the glucose measurements into a vector format. The method can include searching a physical data store by comparing newly received glucose measurements to the centroid of a cluster using a similarity metric. The physical data store can include glucose measurement data and a representation for at least one cluster of glucose measurement data. The representation can approximate a blood glucose profile vector for one cluster of a plurality of glucose profiles segmented by a plurality of time ranges. The method can include classifying newly received glucose measurements by a cluster having a matching similarity metric based on the comparison. The method can include attributing a treatment to the newly received glucose measurement.

[0012] Other features and advantages of the present disclosure will become apparent from the following detailed description when read in conjunction with the accompanying drawings in which like elements are labeled with like numerals.

Brief Description of the Drawings

[0013]

Figure 1A

Figure 1B

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16A

Figure 16B

Figure 16C

Figure 16D

Figure 16E

Figure 16F

Figure 16G

Figure 16H

Figure 16I

Figure 16J

Figure 17

Figure 18

Figure 19

Figure 20

Figure 21

[0014] Embodiments can relate to a system 100 that processes glucose data through efficient glucose database management. The system 100 can include a physical data store 102 that includes glucose measurement data and a representation for at least one cluster of glucose measurement data. The representation can approximate a blood glucose profile vector array for one cluster of a plurality of glucose profiles segmented by a plurality of time ranges. The system 100 can include a processor 104 and a computer memory 106 configured to store instructions 108, which, when executed, cause the processor 104 to perform any of the method steps disclosed herein. The instructions can cause the processor 104 to receive glucose measurement values. The instructions can cause the processor 104 to convert the glucose measurement values into vector form. The instructions can cause the processor 104 to search the physical data store 102 by comparing newly received glucose measurement values to the centroid of a cluster using a similarity metric. The instructions can cause the processor 104 to classify newly received glucose measurement values by a cluster having a matching similarity metric based on the comparison. The instructions can cause the processor 104 to attribute a treatment to newly received glucose measurement values. The treatment can be a command signal, a correction signal, a recommendation, etc. for insulin administration, bolus administration, an exercise routine, a meal consumption routine, a dosing routine, etc.

[0015] The command can cause the processor 104 to store the classification of the newly received glucose measurement in the data store 102 that communicates with one or more other devices 110 (e.g., one or more of a prediction modeling system, a decision support system, an insulin delivery system, an insulin monitoring system, an automatic control system, etc.) configured to use the classification as an input. Additionally, or alternatively, the command can cause the processor 104 to transmit the classification of the newly received glucose measurement to one or more other devices 110 (e.g., one or more of a prediction modeling system, a decision support system, an insulin delivery system, an insulin monitoring system, an automatic control system, etc.) configured to use the classification as an input. Additionally, or alternatively, the command can cause the processor 104 to monitor, analyze, or affect the concentration of glucose levels in the body fluid using the classification of the newly received glucose measurement.

[0016] In some embodiments, the command can cause the processor 104 to receive glucose measurements from a glucose measurement device or data source 112 (e.g., a glucose monitor / sensor, a continuous glucose monitor / sensor, an analysis device, etc.).

[0017] In some embodiments, the system 100 can include a glucose measurement device or data source 112.

[0018] In some embodiments, the system 100 can include a data store 102 that communicates with one or more other devices 110 (e.g., one or more of a prediction modeling system, a decision support system, an insulin delivery system, an insulin monitoring system, an automatic control system, etc.). In some embodiments, the system 100 can include one or more other devices 110 (e.g., one or more of a prediction modeling system, a decision support system, an insulin delivery system, an insulin monitoring system, an automatic control system, etc.).

[0019] In some embodiments, the instructions can cause the processor 104 to calculate the Euclidean distance between one or more newly received glucose measurements and one or more centroids as a similarity metric.

[0020] In some embodiments, the physical data store 102 can include a plurality of clusters. The clusters can be generated by generating an array of glucose measurements for each time range. The plurality of arrays can form a blood glucose profile vector. Weights can be assigned to the arrays. Iterative hierarchical clustering techniques can be applied until one or more clusters are generated that approximate one or more blood glucose profile vectors. One or more clusters of any of the plurality of clusters can be defined by the centroid of the cluster.

[0021] In some embodiments, the iterative hierarchical clustering technique can compute an R 2 value by linear regression on the arrays and vary the weights to maximize the R 2 value.

[0022] In some embodiments, the plurality of time ranges can include five time ranges. For example, the plurality of time ranges can be: 1) Level 2 hypoglycemia below Glucose measurement - 1, 2) Level 1 hypoglycemia within the range from Glucose measurement - 2 to Glucose measurement - 3, 3) Target range (TIR) within the range from Glucose measurement - 4 to Glucose measurement - 5, 4) Level 1 hyperglycemia within the range from Glucose measurement - 6 to Glucose measurement - 7, and 5) Level 2 hyperglycemia above Glucose measurement - 8. In a non - limiting example, Glucose measurement - 1 can be 54 mg / dl, Glucose measurement - 2 can be 54 mg / dl, Glucose measurement - 3 can be 70 mg / dL, Glucose measurement - 4 can be 70 mg / dL, Glucose measurement - 5 can be 180 mg / dL, Glucose measurement - 6 can be 180 mg / dL, Glucose measurement - 7 can be 250 mg / dL, and Glucose measurement - 8 can be 250 mg / dL.

[0023] In some embodiments, a glucose measurement can include a plurality of glucose profiles of an individual. Each glucose profile can include a plurality of glucose measurements obtained over a predetermined period. The instructions can cause the processor 104 to compile the plurality of glucose profiles into a single time series of glucose measurements of the individual. The instructions can cause the processor 104 to classify one or more glucose profiles using one or more clusters to generate a sequence of indices representing the classification of one or more glucose profiles in the single time series of glucose measurements.

[0024] In some embodiments, the instructions cause the processor 104 to generate a trace representing the glucose variability of the individual using the sequence of indices.

[0025] In some embodiments, the instructions can cause the processor 104 to generate an ambulatory glucose profile (AGP) approximated using a sequence of exponents.

[0026] In some embodiments, one or more of the plurality of glucose profiles can be a continuous glucose monitoring (CGM) profile that includes glucose measurements obtained over 24 hours. Additionally, one or more of the plurality of glucose profiles of an individual can be a continuous glucose monitoring (CGM) profile that includes glucose measurements obtained over 24 hours.

[0027] Embodiments can relate to a method of processing glucose data for efficient glucose database management. The method can include receiving glucose measurements. The method can include converting the glucose measurements into a vector format. The method can include searching a physical data store 102 by comparing newly received glucose measurements to the centroid of a cluster using a similarity metric. The physical data store 102 can include glucose measurement data and a representation for at least one cluster of glucose measurement data. The representation can approximate a blood glucose profile vector of one cluster of a plurality of glucose profiles segmented by a plurality of time ranges. The method can include classifying newly received glucose measurements based on the comparison to a cluster having a matching similarity metric. The method can attribute a treatment to the newly received glucose measurement. The treatment can be a command signal, a correction signal, a recommendation, such as an insulin administration, a bolus administration, an exercise routine, a meal consumption routine, a dosing routine, etc.

[0028] In some embodiments, the method can include calculating a Euclidean distance between one or more newly received glucose measurements and one or more centroids as a similarity metric.

[0029] In some embodiments, the physical data store 102 can include: 1) generating an array of glucose measurements for each time range, wherein a plurality of the arrays form a blood glucose profile vector; 2) assigning weights to the arrays; and 3) applying an iterative hierarchical clustering technique that varies the weights until one or more clusters that approximate one or more blood glucose profile vectors are generated; 4) defining one cluster of the set of clusters by the centroid of the cluster.

[0030] The method can include computationally determining an R 2 value by linear regression of the arrays via an iterative hierarchical clustering technique, and varying the weights to maximize the R 2 value.

[0031] As can be appreciated from the present disclosure, embodiments relate to a system 100 and method for processing glucose data with efficient database management. This can be done to classify glucose data and use the classified glucose data to monitor, analyze, and affect the concentration of glucose levels in a body fluid. Some embodiments relate to methods and systems for developing a database for classification, and some embodiments can relate to methods and systems for performing processing using the database.

[0032] Embodiments of system 100 include a processor 104 configured to construct a database of clustering data for glucose measurement classification and / or perform a process for classifying glucose measurements. Processor 104 can be any of the processors 104 disclosed herein. Processor 104 can be part of or communicate with a machine 2000 (logic, one or more components, circuits (e.g., modules), or mechanisms). Processor 104 can be hardware (e.g., a processor, integrated circuit, central processing unit, microprocessor, core processor, computer device, etc.), firmware, software, etc., configured to execute operations by executing instructions implemented in an algorithm, data processing program logic, artificial intelligence programming, automated reasoning programming, etc. Note that the use of processor 104 herein includes any one or combination of a graphics processing unit (GPU), a field programmable gate array (FPGA), a central processing unit (CPU), etc. Processor 104 can include one or more processing modules. The processing module can be a software or firmware operating module configured to perform any of the method steps disclosed herein. The processing module can be implemented as software and stored in a memory, which is operably associated with processor 104. The processing module can be implemented as a web application, a desktop application, a console application, etc. Exemplary embodiments of processor 104 and machine 2000 are described below.

[0033] Processor 104 can include or be associated with a computer or machine-readable medium 2002. As described in detail below, the computer or machine-readable medium 2002 can include a memory 106. Any of the memories 106 described herein can be a computer-readable memory configured to store data. The memory 106 can include volatile or non-volatile, temporary or non-temporary memory and can be implemented as in-memory, active memory, cloud memory, etc. Embodiments of the memory 106 can include a processor module and other circuitry that enable the transfer of data between the memory 106, which can include other components of the communication system's round-trip. This transfer can be via hardwire or wireless transmission. The communication system can be used in combination with switches, receivers, transmitters, routers, gateways, waveguides, etc., and can include a transceiver that facilitates communication with any other component or combination of components of the communication system via a communication method or protocol for controlled and coordinated signal transmission and processing. Transmission can be via a communication link. The communication link can be electronic-based, optical-based, optical-electronic-based, quantum-based, etc.

[0034] The computer or machine-readable medium 2002 can be configured to store one or more instructions 108. The instructions 108 can be in the form of an algorithm, program logic, etc. that cause the processor 104 to build and / or implement a classification model.

[0035] Processor 104 can communicate with other processors of other devices 110 (e.g., predictive modeling systems, decision support systems, insulin delivery systems, insulin recommendation systems, blood glucose or insulin monitoring systems, glucose or insulin management systems, automatic control systems, etc.) configured to use classification as an input. Any of these other devices 110 can include any of the exemplary processors disclosed herein. Any of the processors can have a transceiver or other communication device / circuit to facilitate the transmission and reception of wireless signals. Any of the processors can include an application programming interface (API) as software means to enable two applications to talk to each other. The use of the API enables the software of processor 104 of system 100 to communicate with the software of the processors of other devices 110.

[0036] Any transmission between processors / devices / systems / modules can be a push, operation, pull operation, or a combination of both. Any transmission can be a direct transmission between two components or a transmission via means. The means can be, for example, a memory, database, data store, etc. For example, data from one processor can be sent to a database for storage before being sent to another processor. As another example, data can be sent to an intermediate processor or processing module to process, format, encode the data, etc. before being sent to another processor. The data transmission between components can be performed continuously, periodically, or at some other predetermined schedule as required by a control signal based on conditions satisfied with respect to the algorithm function.

[0037] Exemplary Systems and Methods for Developing a Database of Clustered Datasets Embodiments can relate to a system 100 for developing a database for classifying glucose data. The system 100 can include a processor 104. The system 100 can include a computer memory 106 storing instructions 108 that, when executed, cause the processor 104 to perform any of the method steps disclosed herein. The instructions 108 can cause the processor 104 to receive glucose profile data. The glucose profile data can include one or more glucose measurements. The glucose measurements can be time series measurements representing a glucose level profile (e.g., pattern, behavior, trend, etc.). The glucose profile data can be historical, current, and / or real-time data. The glucose profile data is received by the processor 104. This can be done continuously, periodically, or according to some other predetermined schedule. The glucose profile data can be pulled by the processor 104 from a data source 112 and / or pushed from the data source 112 to the processor 104. The data source 112 can be a device that generates glucose measurements (e.g., a glucose monitor / sensor, a continuous glucose monitor / sensor, an analysis device, etc.) or a data store 102 (e.g., a database) that stores glucose profile data. The glucose measurements can be a liquid such as a body fluid. The processor 104 can store the glucose profile data in a temporary or persistent memory for later processing or process the glucose profile data when received. For example, the processor 104 can receive the glucose profile data and aggregate the glucose profile data in storage. Aggregation can be performed, for example, based on metadata, the type of data, what the data represents, the time the data was received, the time the data was generated, etc.

[0038] Command 108 can cause processor 104 to generate a set of clusters from glucose profile data. As a non-limiting example, the command can cause processor 104 to perform machine learning data mining techniques that divide groups of objects in the glucose profile data into classes of similar objects. The clusters can be configured to approximate multiple time ranges of the glucose profile data. For example, one or more of the clusters can be clustered to approximate one or more time ranges in the glucose profile data. The time range can be the duration that the glucose measurements in the glucose profile data have values within one range of the glucose measurements. For example, there can be a duration in which the glucose profile data has glucose measurements of G1, and G1 falls within the glucose measurement range x-y. A plurality of time range sets can exist. In other words, it can be advantageous to know the time when the glucose profile data had glucose measurements that fell within a predetermined time range and how many glucose measurements had glucose measurements that fell within the predetermined time range. This information can be used to generate clusters that represent the same thing.

[0039] Command 108 can cause processor 104 to generate a set of one or more clusters. Any of the sets of clusters can be generated using hierarchical clustering techniques. For example, a set of clusters can be generated by generating an array of glucose measurements for each time range. One or more arrays can form a vector. For example, if there are five time ranges, the five time ranges will generate five arrays. More or fewer time ranges (and arrays) can also be used. One or more arrays (e.g., all five arrays) can be used to generate a vector. Since the arrays contain glucose measurements, the vector can be a blood glucose profile vector (or one or more blood glucose profile vectors). A weighting can be assigned to the arrays, and the step of assigning the weighting to one or more arrays can be included. The weighting can be, for example, a value from 0 to 1, a weighting function, or any other mathematical operator that gives a desired influence or effect to the arrays. Any one or combination of the weightings can be determined by an optimization function, an objective function, a cost function, etc. Any one or combination of the weightings can be fixed or variable. The weighting can be variable and randomly set to any value for the first or initial iteration. The weighting can be changed for each iteration until an optimum is reached. For example, command 108 can cause processor 104 to apply an iterative hierarchical clustering technique that changes the weighting until a set of clusters that approximates the blood glucose profile vector is generated. This approximation can be a best approximation, a desired approximation, an optimum approximation, etc. For example, the optimum approximation can be an approximation defined by an optimization function, an objective function, a cost function, etc.

[0040] Command 108 can cause the processor 104 to define one cluster (which can include any number of clusters) of the cluster set by the centroid of the cluster. For example, an individual cluster of the cluster set can be defined by the centroid of that individual cluster. This can be done for one or more of the clusters. The centroid can be a statistical (weighted or unweighted) center, average, mode, etc. Thus, each cluster can be defined by a value or variable representing the centroid of the cluster. The cluster set can be a set of values or variables representing the clusters. As shown herein, each value or variable represents a time range or is an approximation of a time range of glucose profile data. It should be noted that the cluster data is an exact representation (or proxy) for the time range of the glucose data profile but is a significantly reduced data set.

[0041] As will be appreciated, system 100 can reduce the data required for glucose analysis and reduce the computational resources, etc. required for a system to process such data. For example, instead of transmitting / processing a daily continuous glucose monitoring (CGM) profile (which is typically 288 data points), a single number can be transmitted / processed. System 100 can generate cluster data from data from any type of glucose measurement system (e.g., data from any type of measurement system, data from separate glucose measurement systems, non-standardized data, etc.) and data related to one or more of type 1 diabetes, type 2 diabetes, etc. In other words, system 100 can be agnostic to the type of data, the pattern of measurements, etc. System 100 improves the robustness and accuracy in glucose profile data except where it is considered inappropriate due to lost data, data from separate data sources, or non-standardized data, etc. that may be used to generate the clusters.

[0042] Command 108 can cause processor 104 to store the cluster set in data store 102. This can be a physical data store 102. Data store 102 can communicate with other devices 110 (e.g., one or more of a predictive modeling system, a decision support system, an insulin delivery system, an insulin monitoring system, an automatic control system, etc.) configured to use one or more clusters of the cluster set as input. For example, as will be described in more detail below, a decision support system can compare new glucose profile data to a model, classify the new glucose profile data as falling within one or more of the time ranges, and do this by assigning one or more centroid values to the new glucose profile data. For example, the new glucose profile data can be matched to one or more clusters and the centroid cluster value can be given to the matched cluster. Alternatively, processor 104 can perform this function and send the value to the decision support system. This value can then be used as a proxy or surrogate for the time range of the glucose measurements of the glucose profile. Data store 102 can be part of system 100 or part of another system. Additionally, or alternatively, system 100 can be a predictive modeling system, a decision support system, an insulin delivery system, an insulin monitoring system, an automatic control system, etc., and can use the data directly (e.g., preventing the use of data store 102). The cluster set can be stored or used as a classification database. This database can be modified, learned, etc., based on additional or updated data.

[0043] Glucose profile data can include a plurality of glucose profiles. System 100 can classify one or more of the plurality of glucose profiles by one or more clusters of a cluster set. One or more of the glucose profiles can include a plurality of glucose measurements taken over a predetermined period. For example, one or more of the glucose profiles can be a continuous glucose monitoring (CGM) profile. One or more of the CGM profiles can include glucose measurements taken over a 24-hour period, and can include glucose measurements taken every 5 minutes over a 24-hour period. As shown herein, System 100 can manipulate missing data. Thus, if each CGM profile data is expected to include 288 glucose measurements (e.g., 288 data points), System 100 can generate useful clusters with less effort.

[0044] Instruction 108 can cause the processor 104 to compute an R value by linear regression for each array when implementing an iterative hierarchical clustering technique. 2 Instruction 108 can cause the processor 104 to change one or more weightings to maximize the R value. The change in weighting can be through iterative or recursive processing and can be managed by an optimization function, objective function, cost function, etc. 2 When other numbers and other ranges of time-in-range can be used, the plurality of time-in-ranges shall be assumed to include five time-in-ranges. These are shown below.

[0045] Level 2 hypoglycemia below glucose measurement -1; Level 1 hypoglycemia within the range from glucose measurement -2 to glucose measurement -3; Target range (TIR) within the range from glucose measurement -4 to glucose measurement -5; Level 1 hyperglycemia within the range from glucose measurement -6 to glucose measurement -7; and ​Level 2 hyperglycemia exceeding the glucose measurement value - 8.

[0046] The time in range is shown below. The glucose measurement value - 1 is 54 mg / dl; The glucose measurement value - 2 is 54 mg / dl; The glucose measurement value - 3 is 70 mg / dL; The glucose measurement value - 4 is 70 mg / dL; The glucose measurement value - 5 is 180 mg / dL; The glucose measurement value - 6 is 180 mg / dL; The glucose measurement value - 7 is 250 mg / dL; and The glucose measurement value - 8 is 250 mg / dL.

[0047] When developing a database, glucose profile data can include glucose measurements from one or more individuals. One or more glucose profiles can exist for each individual. This robust dataset enables the use of models to perform multivariate analyses regarding determining blood glucose trends, predicting blood glucose states, and conditions and factors resulting from or related to blood glucose states (e.g., eating behavior, exercise behavior, medical conditions, age, gender, race, heart rate, respiratory rate, blood oxygen saturation, etc.). For example, multivariate modeling techniques can be used to determine conditions or factors that statistically contribute to changes in blood glucose states, changes in the risk of hypoglycemia or hyperglycemia, etc., which can also be used to estimate the probabilities of similar things. Multivariate modeling techniques can include one or more of logistic regression with or without cubic splines, random forest, xgboost, support vector machine, nearest neighbor method, artificial neural network, and / or long short-term memory (LSTM), multivariate analysis of variance (MANOVA), multivariate analysis of covariance (MANCOVA), principal component analysis (PCA), canonical correlation analysis, redundancy analysis (RDA), correspondence analysis (CA), canonical correspondence analysis (CCA), multidimensional scaling, discriminant analysis, linear discriminant analysis (LDA), clustering system, recursive adaptive partitioning, vector autoregression, principle response curve analysis (PRC), etc. As a non-limiting example, the mean, standard deviation, and / or cross-correlation, one or more of the conditions or factors, and cluster centroids can be fit to a logistic ridge regression model using cubic splines, for example, to generate an output that is an estimate or probability of the occurrence of a blood glucose state.

[0048] As described above, one or more glucose profiles for each individual can be present. When multiple glucose profiles exist for an individual, instruction 108 can cause processor 104 to compile the multiple glucose profiles into a single time series of glucose measurements for the individual, e.g., a single time series of glucose measurements spanning the entire set of glucose profiles for this individual. Instruction 108 can cause processor 104 to classify each glucose profile by one or more clusters of a cluster set and generate a sequence of indices representing the classification of each glucose profile into the single time series of glucose measurements. Instruction 108 can cause processor 104 to store the sequence of indices in data store 102, which is part of a database. This can be done for one or more individuals. Thus, data store 102 can have a sequence of indices for each individual, and each sequence is an approximation of the time-in-range of glucose measurements in its respective time series. Note that there can be one or more time series of data for an individual. Also, one or more sequences of indices can exist for any single time series of data.

[0049] The sequence of indices enables the use of the database to perform multivariate analysis regarding determining glucose trends, predicting glucose states, and conditions and factors that cause or are related to an individual's glucose state (e.g., eating behavior, exercise behavior, medical conditions, age, gender, race, heart rate, respiratory rate, blood oxygen concentration, etc.). Additionally, or alternatively, instruction 108 can cause processor 104 to use the sequence of indices to generate a trace representing an individual's glucose variability. This can be done for one or more individuals. Also, one or more traces can be present for an individual. Instruction 108 can cause processor 104 to store the trace in data store 102, which is part of a database.

[0050] Exemplary System and Method for Classifying Glucose Data Embodiments can relate to a system 100 for classifying glucose data. The system 100 can be configured to implement embodiments of the methods disclosed herein that use a database of clustered data to classify glucose data. The system 100 can include a processor 104. The system 100 can include a computer memory 106 that stores instructions 108 that, when executed, cause the processor 104 to implement or apply embodiments of the methods disclosed herein. The instructions 108 can cause the processor 104 to receive glucose profile data that includes a plurality of glucose measurements. The glucose data is assumed to be for a single individual so that the individual's blood glucose status can be evaluated or determined by comparing the individual's glucose profile data to the clusters of the database. However, the glucose profile data can be for one or more individuals. The glucose profile data is considered to be up-to-date data (e.g., data collected in real time or within the past 24 hours), but the glucose profile data can be historical, current, and / or real-time data.

[0051] Command 108 can cause the processor 104 to classify glucose profile data or a portion thereof by comparing the glucose profile data to an embodiment of a database. The database can include a set of clusters configured to approximate one or more blood glucose profile vectors for an individual and / or a group of individuals to which the individual belongs (e.g., individuals are grouped by age, gender, race, medical condition, etc.). A blood glucose profile vector is an array of previously processed glucose profile data segmented by a plurality of time-in-ranges. The previously processed glucose profile data includes historical glucose data, but can also include current or real-time glucose data. The previously processed glucose data can be personal data, data of an individual within a group of individuals (which may or may not include personal data), data of an individual that can or cannot be within a group of individuals, etc. One or more of the time-in-ranges can be the duration that the glucose measurements of the previously processed glucose profile data had values within a range of glucose measurements.

[0052] Command 108 can cause the data store 102, which communicates with the processor 104 and other devices 110 (e.g., one or more of a predictive modeling system, a decision support system, an insulin delivery system, an insulin monitoring system, an automated control system, etc.) configured to use the classification as an input, to store the classification of the glucose profile data. Additionally or alternatively, command 108 can cause the processor 104 to transmit the classification of the glucose profile data to other devices 110 (e.g., one or more of a decision support system, an insulin delivery system, an insulin monitoring system, etc.) configured to use the classification as an input. Additionally or alternatively, command 108 can cause the processor 104 to monitor, analyze, and / or manipulate the concentration of glucose levels in a body fluid using the classification.

[0053] Command 108 can cause the processor 104 to classify glucose profile data by comparing the glucose profile data to the centroid of a set of clusters. For example, glucose profile data that matches exactly, approximately, or similarly to the centroid can be classified as having an exact, approximate, or similar time range pattern as the cluster to which the centroid belongs. The glucose profile data of any one individual can have one or more classifications. The glucose profile data of any one individual can include one or more glucose profiles of the individual. One or more classifications can be present in any one glucose profile. The glucose profile to be classified is considered to have only one classification (i.e., best matches the centroid of one cluster). In the event of a low likelihood of a close similarity score, the first match can be selected.

[0054] The model can have a set of clusters. The number of clusters can be, for example, 35. More or fewer clusters per set can also be used. One or more sets of clusters can be present. The number of clusters in one set can be the same as or different from the number of clusters in another set. The number of clusters, the number of sets, etc. can be set according to desired design conditions (e.g., optimization, computational resources, processing speed, accuracy, robustness, etc.). The comparison can be made between the glucose profile data and one or more clusters (or centroids) within the same set, different sets, a single set of clusters, multiple sets of clusters, etc.

[0055] Command 108 can cause the processor 104 to compare the glucose profile data with one or more centroids using a similarity measure. The glucose profile data can be classified using the cluster with the best similarity measure. The similarity measure can be a numerical value that falls within a range of values (e.g., from 0 to 1). A similarity measure of 0 can indicate a match, while a similarity measure of 1 can indicate a mismatch with a gradation of the degree of matching between 0 and 1. Alternatively, a similarity measure of 1 can indicate a match, while a similarity measure of 0 can indicate a mismatch with a gradation of the degree of matching between 1 and 0. Other similarity measure schemes can also be used. In an exemplary embodiment, command 108 can cause the processor 104 to calculate the Euclidean distance between one or more glucose profile data points and one or more centroids as the similarity measure. The distance can be normalized, for example, to fit within a range from 0 to 1.

[0056] Glucose profile data can include one or more glucose profiles of an individual. Each glucose profile can include a plurality of glucose measurements taken over a predetermined period (e.g., over 24 hours). Instruction 108 can cause processor 104 to compile a plurality of glucose profiles of an individual into a single time series of glucose measurements. Instruction 108 can cause processor 104 to classify each glucose profile by one or more clusters of a cluster set and generate a sequence of indices representing the classification of each glucose profile into the single time series of glucose measurements. Instruction 108 can cause processor 104 to store the sequence of indices in a data store 102 that communicates with one or more other devices 110 (e.g., one or more of a predictive modeling system, a decision support system, an insulin delivery system, an insulin monitoring system, an automatic control system, etc.) configured to use the classification as an input. Additionally, or alternatively, instruction 108 can cause processor 104 to transmit the sequence of indices to one or more other devices 110 (e.g., one or more of a predictive modeling system, a decision support system, an insulin delivery system, an insulin monitoring system, an automatic control system, etc.) configured to use the classification as an input. Additionally, or alternatively, instruction 108 can cause processor 104 to monitor, analyze, and / or manipulate the concentration of glucose levels in a body fluid using the classification.

[0057] Instruction 108 can cause processor 104 to generate an approximate ambulatory glucose profile (AGP) using the sequence of indices. Additionally, or alternatively, one or more other devices 110 (e.g., one or more of a predictive modeling system, a decision support system, an insulin delivery system, an insulin monitoring system, an automatic control system, etc.) can generate an approximate ambulatory glucose profile (AGP) using the sequence of indices.

[0058] As shown in this specification, glucose profile data can include a plurality of glucose profiles, and each glucose profile includes a plurality of glucose measurements obtained over a predetermined period. For example, each glucose profile can be a continuous glucose monitoring (CGM) profile that includes glucose measurements obtained over 24 hours.

[0059] If other numbers and other ranges of time in range can be used, the plurality of time in ranges shall include five time in ranges. These are as follows. Level 2 hypoglycemia below glucose measurement -1; Level 1 hypoglycemia within the range from glucose measurement -2 to glucose measurement -3; Target range (TIR) within the range from glucose measurement -4 to glucose measurement -5; Level 1 hyperglycemia within the range from glucose measurement -6 to glucose measurement -7; and Level 2 hyperglycemia above glucose measurement -8.

[0060] The time in range is shown below. Glucose measurement -1 is 54 mg / dl; Glucose measurement -2 is 54 mg / dl; Glucose measurement -3 is 70 mg / dL; Glucose measurement -4 is 70 mg / dL; Glucose measurement -5 is 180 mg / dL; Glucose measurement -6 is 180 mg / dL; Glucose measurement -7 is 250 mg / dL; Glucose measurement -8 is 250 mg / dL.

[0061] Any of the other devices 110 can be configured to generate an output based on the classification input. In some embodiments, the other devices 110 can be part of the system 100. For example, the system 100 can include the other devices 110. The classification can be used by the processor 104 or a processor of the other device 110 to generate a signal. a. Recommendation or implementation of a process to obtain additional data (e.g., signals are generated that request additional patient data, insulin delivery data, metabolic data, etc.); b. Recommendation or implementation of a process to initiate preventive or mitigating means (e.g., signals are generated that modify the insulin rate, modify behavior, etc.); c. Recommendation or implementation of a process to initiate extended monitoring (e.g., signals are generated that inform the user that the risk of hypoglycemia is increasing and that additional monitoring should be done). d. This is an alert signal or command signal to an insulin delivery device that modifies, for example, the insulin rate or dosage.

[0062] Either the system 100 or any of the other devices 110 can include a display configured to generate a user interface. The user can control aspects of the system 110 via the user interface. Additionally, the user interface can generate a graphics display that displays aspects of the classification and other outputs, audible, graphic, or text alerts, etc.

[0063] System 100 can include a processor 104 in combination with one or more data stores 120. The data store 102 can be configured to include a plurality of classification databases. For example, system 100 can be configured to generate a plurality of classification databases. The processor 104 can be configured to use any one or a combination of the plurality of classification databases. Each classification database can be generated based on available glucose and other patient data, the expected availability of glucose or other patient data, the quality of glucose or other patient data (how reliable the data is), the frequency of glucose or other patient data (how frequently it is generated or available), the dimensions of glucose or other patient data (the number of attributes or variables the data has), etc. For example, the first classification database can be generated for a data set where certain types of data are sparse while other types of data are abundant, and the second classification database can be generated for a data set where the reliability of certain data is low while it is high for other types of data. The types of patient data can include from the data source 112 from which the data is received or reception is attempted (or requested), the attributes included in the data, the number of attributes the data has, etc. The classification databases can be generated for an expected data flow, thereby generating a plurality of classification databases. The plurality of classification databases can be stored in one or more data stores 102. The processor 104 can communicate with the data store 102 to access any one or a combination of the plurality of classification databases.

[0064] The processor 104 can be configured to switch the implementation from the first classification database to the second classification database based on at least one or more of the type of data, the availability of the data, the reliability of the data, etc. The processor 104 can detect a change (e.g., based on metadata) and switch the classification database.

[0065] Processor 104 can be configured to update the classification database based on new data. As described above, the glucose profile data can be historical, current, and / or real-time data, and can be received continuously, periodically, or on some other predetermined schedule, and can include information regarding glucose episodes, treatments, and the like. System 100 can update any one or combination of the classification databases based on the updated data. The updated classification database can replace the existing classification database in data store 102. Alternatively, the updated classification database can be added to the plurality of classification databases if the updated classification database is sufficiently different or more suitable for the patient data scenario than any other existing classification database.

[0066] As will be appreciated from the present disclosure, aspects of embodiments of the present invention provide, among other things, a system, method, and computer-readable medium for identifying clinical similarity clusters of daily continuous glucose monitoring (CGM) profiles.

[0067] Aspects of embodiments of the present invention provide, among other things, a system, method, and computer-readable medium that perform the steps of: a) constructing and then fixing a set of clinical similarity clusters (CSCs) based on the property that for any other daily continuous glucose monitoring (CGM) profile, there exists a clinical similarity cluster (CSC) that approximates the time-in-range of the daily CGM profile; and b) determining the approximation of any daily CGM profile by the CSCs.

[0068] Aspects of embodiments of the systems, methods, and computer-readable media of the present invention include, for example, but are not limited to, two steps. For example, the first step can include constructing and then fixing a set of clinically similar clusters (CSCs) based on the property that for any other daily CGM profile, there exists a clinically similar cluster (CSC) that approximates the time range of the daily CGM profile, thereby preserving important clinically relevant features of the daily CGM profile.

[0069] The set can be defined using hierarchical clustering, where the weights of the input columns are varied until the set of CSCs has the desired performance when approximating the time range of the daily CGM profile. The second step can include, for example, computationally calculating a similarity measure (e.g., Euclidean distance) between a candidate daily CGM profile and the centroid of each CSC, and selecting a single CSC with the minimum similarity measure value, which can include determining an approximation of any daily continuous glucose monitoring (CGM) profile by the CSC.

[0070] In one embodiment, when these steps are achieved, any daily CGM profile can be mapped to a CSC, and then the sequence of an individual's CSCs can be used as a surrogate for the individual's ambulatory glucose profile (AGP) and the associated time range of the original daily CGM profile. In addition to the AGP and associated metrics, the sequence of CSCs provides information regarding the timing and day-to-day variability of the patient's clinically relevant glucose events. Potential applications of aspects of embodiments of the systems, methods, and computer-readable media of the present invention include, for example, but are not limited to, one or more of the following: (i) data construction and dimensionality reduction, (ii) database indexing, (iii) compression / encryption of daily CGM profiles, (iv) discrimination between health states and treatment modalities, (v) CGM exchange for common clinical tests, (vi) CGM pattern recognition and prediction, or (vii) tracking of disease progression.

[0071] Aspects of embodiments of the systems, methods, and computer-readable media of the present invention can be configured to work, among other things, with daily CGM profiles generated by sensors having various sampling resolutions, and daily CGM profiles having missing data up to a certain threshold. One important advantage of aspects of embodiments of the present invention is, but is not limited to, the ability to classify all CGM profiles into relatively small, finite, and fixed ones across patient groups and overall health states, and the set of CSCs accurately represents the clinical status of these patients. Aspects of embodiments of the present invention also add, but are not limited to, time-varying components to commonly accepted CGM data representations such as the AGP and the time-in-range associated with the AGP.

[0072] Aspects of embodiments of the present invention provide, among other things, systems, methods, and computer-readable media for identifying clinical similarity clusters of daily CGM profiles.

[0073] Aspects of embodiments of the present invention provide, among other things, systems, methods, and computer-readable media for classifying and clinically interpreting daily CGM profiles.

[0074] Aspects of embodiments of the present invention provide, among other things, systems, methods, and computer-readable media for defining a set of CGMs, and any daily CGM profile can be classified as one of the CSCs, and the CSCs reliably approximate the clinical characteristics of the daily CGM profiles.

[0075] Aspects of embodiments of the present invention provide, among other things, systems, methods, and computer-readable media for clearly distinguishing between health states and treatment modalities and providing CSCs used as representations of the rate of change of a person's blood glucose over time.

[0076] One aspect relates to a method for identifying clinical similarity clusters of daily continuous glucose monitoring (CGM) profiles as described herein. The method comprises A step of obtaining individual i, wherein each individual i has a single CGM time series generated during the research in which the individual participates A sequence s of indices (discontinuous if possible) indicating the CSC into which each daily CGM profile is classified i A step of classifying all of the daily CGM profiles into a single time series of individuals resulting in where each entry s of the sequence i corresponds to a single day of observation, the days being ordered by the date of occurrence, and f i (t) is the number of unique CSCs visited by individual i after t days of observation, and f i is a trace of the number of unique CSCs visited by individual i over time, providing an ideal of variability in the individual's blood glucose, a step A step of providing an average trace that indicates the average behavior of individuals in a subgroup and helps to highlight differences in behavior between different subgroups where, for a given subgroup G, the average value is TIFF2025516193000002.tif14150, where TIFF2025516193000003.tif6150 is the total number of individuals in the subgroup, where all individuals in the subgroup do not have the same number of days observed in the sequence of CSC indices, and time TIFF2025516193000005.tif6150 is defined only if there is a minimum number of sequences of TIFF2025516193000006.tif6150, the minimum number of sequences being a function of the subgroup TIFF2025516193000007.tif6150, a step A step of distinguishing a health state and a treatment mode because the health state and the treatment mode track a blood glucose variability rate, wherein on average, healthy individuals have the lowest blood glucose variability rate, followed by individuals with T2D and then individuals with T1D, and for individuals with T1D, individuals using MDI as an average treatment mode have the highest blood glucose variability rate, followed by individuals in PMP and then individuals in CLC. It can include.

[0077] The method further A step of performing data construction and dimensionality reduction, wherein all possible daily CGM profiles clinically represented by AGP and this time-in-range are reduced to a finite and fixed set of CSCs. A step of database indexing, wherein the database is indexed by a structure defined by the CSCs that guarantees a fast and efficient search of a subgroup of similar daily CGM profiles. A step of compressing and / or encrypting the daily CGM profile. A step of distinguishing between a health state and a treatment mode. It can include. The ability of the CSCs to distinguish between health states with high fidelity serves as a replacement for a clinical test of wearing a CGM for a specified number of days in a home environment with a predefined diet schedule and physical activity that will result in a diagnostic outcome.

[0078] One aspect relates to a method of performing the following steps: a) constructing and then fixing a set of clinical similar clusters (CSCs) by virtue of the property that there exists a clinical similar cluster (CSC) that approximates the time-in-range of the daily CGM profile for any other daily continuous glucose monitoring (CGM) profile, and b) determining the approximation of any daily CGM profile by the CSCs, as described herein.

[0079] One aspect relates to a system that performs the following steps, as described herein: a) constructing and then fixing a set of clinical similarity clusters (CSCs) based on the property that for any other daily continuous glucose monitoring (CGM) profile, there exists a clinical similarity cluster (CSC) that approximates the time-in-range of the daily CGM profile; and b) determining an approximation of any daily CGM profile by the CSCs.

[0080] One aspect relates to a computer-readable storage medium storing computer-executable instructions that, when executed by one or more processors, cause one or more computers to perform functions that include: a) constructing and then fixing a set of clinical similarity clusters (CSCs) based on the property that for any other daily continuous glucose monitoring (CGM) profile, there exists a clinical similarity cluster (CSC) that approximates the time-in-range of the daily CGM profile; and b) determining an approximation of any daily CGM profile by the CSCs, as described herein.

[0081] One aspect relates to a method configured to provide a two-step iterative process for identifying a fixed set of clinical similarity clusters (CSCs) for daily CGM profiles. The two-step process uses hierarchical clustering in a training data set configured to identify a candidate set of CSCs. The two-step process uses a validation data set configured to evaluate the performance of the candidate set of CSCs. The ability of the CSCs to reliably capture five different time-in-ranges of the daily CGM profiles to be classified is evaluated. Next, a fixed set of 35 CSCs TIFF2025516193000008.tif6150 was used to classify daily CGM profiles in another test data set, where the results showed that the set was robust and properly generalized.

[0082] In some embodiments, the distribution of daily CGM profiles to different CSCs has been shown to be unique to health status and treatment modality.

[0083] One aspect relates to a method of visualizing an individual's blood glucose control. Using clinical similarity clusters (CSCs), differences in blood glucose control among individuals with the same health status and treatment modality can be visualized to identify individuals who may require more individualized attention. In some embodiments, is TIFF2025516193000009.tif8150, TIFF2025516193000010.tif6150 is for an individual TIFF2025516193000011.tif6150 of TIFF2025516193000012.tif6150 is the number of unique CSCs required to classify the daily CGM profile.

[0084] In some embodiments, the total number of unique CSCs can be bounded (i.e., there are only 35 different CSCs), if TIFF2025516193000013.tif6150 is large, then TIFF2025516193000014.tif6150 tends to be zero, TIFF2025516193000015.tif6150 is fixed to 28 daily profiles (i.e., 4 weeks of data), TIFF2025516193000016.tif6150 is TIFF2025516193000017.tif6150 = the average value of the values generated by computer calculation of each TIFF2025516193000018.tif6150 using a sliding window of 28 daily CGM profiles Defined as TIFF2025516193000019.tif6150, where the sliding window advances every 7 days (1 week of data), and each sliding window TIFF2025516193000020.tif6150 must have at least a 14-day CGM profile (2 weeks of data) for computer calculation.

[0085] In some embodiments, TIFF2025516193000021.tif6150 is the individual of TIFF2025516193000022.tif6150 TIFF2025516193000023.tif6150 is the average CSC index of the daily CGM profile, and TIFF2025516193000024.tif6150 is TIFF2025516193000025.tif6150 = the average value of the values generated by computer calculating each TIFF2025516193000026.tif6150 using a sliding window of the 28-day CGM profile, and TIFF2025516193000027.tif6150 is set as such, where the sliding window advances every 7 days, and each sliding window TIFF2025516193000028.tif6150 must have at least a 14-day CGM profile for computer calculation.

[0086] In some embodiments, any daily CGM profile can be approximated by one of 35 pre-fixed clinical similarity clusters (or a specified number of pre-fixed clinical similarity clusters). The approximation means that when the daily CGM profile is classified into a CSC, the CSC preserves the information carried by the original daily CGM profile in terms of the time-in-range scale system. The CSC is expanded and completed to a certain extent, and the interpretation of the CGM data is provided by the AGP / TIR system, where AGP / TIR is a static snapshot of 14 days (or a specified number of days) of data, and the sequence of CSCs obtained from the same data tracks the progression of blood glucose control over time. The time series of CSCs over 14 days (or a specified number of days) indicates how stable or unstable a person's blood glucose control is.

[0087] Aspects of embodiments of the systems, methods, and computer-readable media of the present invention generally relate to, but are not limited to, pharmaceuticals and medical devices used in the insulin treatment of diabetes and other metabolic disorders, including type 1 and type 2 diabetes, type 2 (TID, T2D), latent autoimmune diabetes in adults (LADA), postprandial or reactive hyperglycemia, or insulin resistance. In alternative embodiments, aspects of embodiments of the present invention define a set of clinically similar clusters (CSCs) based on the property that for any other daily CGM profile, there exists a clinically similar cluster (CSC) that approximates the time-in-range of the daily CGM profile, and then fixes this, thereby preserving important clinically relevant features of the daily CGM profile. When the CSCs are defined and fixed, any daily CGM profile can be mapped to the CSCs, and then the sequence of an individual's CSCs can be used as a surrogate for the individual's ambulatory glucose profile (AGP) and the associated time-in-range of the original daily CGM profile. In addition to the AGP and associated metrics, the sequence of CSCs provides information regarding the timing and day-to-day variability of the patient's clinically relevant glucose events. One important advantage of aspects of embodiments of the present invention is the ability to classify all CGM daily profiles into a relatively small finite and fixed number across patient groups and health states, and the set of CSCs accurately represents the clinical status of these patients.

Example

[0088] The following are examples for developing, testing, and implementing the systems and methods of the embodiments disclosed herein. The following are merely exemplary and should not be construed in a limiting sense.

[0089] Example 1 Data The data used in this work was generated from the following. 1. University of Virginia Diabetes Technology Center (i.e., DCLP1

[19] , DCLP3

[20] , DIAMOND1

[21] , DIAMOND2

[22] , DSS1

[23] , NIGHTLIGHT

[24] , and TRIALNET study) and 2. Website of the Health Research Jæbcentret (CITY

[25] , DCLP5

[26] , NDIAB

[27] , MDEX

[28] , REPLACE - BG

[29] , RT - CGM

[30] , SENCE

[31] , SEVHYPO

[32] , and WISDM

[33] studies).

[0090] From each of these studies, the procedure outlined in paragraph III.A of

[34] was used to process the CGM time series and define the daily CGM profile, where the daily CGM profile is a time series of 288 blood glucose data points collected every 5 minutes during a midnight - to - midnight (24 - hour) period. 204,710 daily CGM profiles from these 16 different studies were used to form three different datasets, each with a distinct purpose. 1. Training dataset - This dataset was composed of 23,916 daily CGM profiles taken from the DCLP1, DCLP3, DIAMOND1, DIAMOND2, DSS1, and NIGHTLIGHT studies and was used to define a candidate set of CSCs. 2. Validation dataset - This dataset was composed of 37,758 daily CGM profiles taken from the DCLP1, DCLP3, DIAMOND1, DIAMOND2, DSS1, and NIGHTLIGHT studies and was used to a) evaluate the performance of each candidate set of CSCs and b) select the final and fixed set of CSCs. 3. Test dataset - This dataset was composed of 143,036 daily CGM profiles taken from the CITY, DCLP5, DIAMOND2, NDIAB, MDEX, REPLACE - BG, RT - CGM, SENCE, SEVHYPO, TRIALNET, and WISDEM studies and was used to evaluate the robustness and generalizability of the final selected set of CSCs.

[0091] The study represents healthy individuals, individuals with type 1 diabetes (T1D), and individuals with type 2 diabetes (T2D). The study also represents a variety of treatment modalities including multiple daily injections (MDI), insulin pumps (PMP), and closed-loop control (CLC).

[0092] From 16 datasets, 2,462 subjects and a total of 204,710 daily CGM profiles were collected. The characteristics of the participants in each study are detailed in Table 1. Table 1 - Characteristics of the 16 datasets used in this work. Statistical values are shown as mean (SD) unless otherwise indicated. TID is type 1 diabetes, T2D is type 2 diabetes, BMI is body mass index, CGM is continuous glucose monitoring, MDI is multiple daily injections, PMP is insulin pump, and CLC is closed-loop control. * indicates that data was not available at the subject level and was taken from the study protocol. TIFF2025516193000029.tif249155

[0093] Many (95.9%) of the daily CGM profiles were generated by subjects with T1D being treated by either MDI, insulin pump (PMP), or closed-loop control (CLC). Two datasets focused on children with T1D (DCLP5 and SENCE). The DIAMOND2 dataset included data from participants with T2D on MDI treatment and represented 3.6% of the generated daily GCM profiles. The NDIAB and TRIALNET datasets included data from non-diabetic (healthy) individuals. Glycemic control as evaluated by the mean HbA1c of each participant at baseline ranged from 5.2% (NDIAB study) to 9.1% (CITY study) for non-diabetic participants. Healthy individuals generally had data for less than 7 days, while people in most other studies had data for an average of 5 or more weeks.

[0094] Identification of Clinically Similar Clusters of Daily CGM Profiles In one embodiment, for example, the 5 time-in ranges are level 2 hypoglycemia [T54]: 54 mg / dl, level 1 hypoglycemia [T70]: 54 - 69 mg / dL, within the target range [TIR]: 70 - 180 mg / dL, level 1 hyperglycemia [T180]: 180 - 250 mg / dL, and level 2 hyperglycemia [T250]: above 250 mg / dL during a 24-hour period. In other embodiments, other time-in ranges can be used depending on the specific clinical metrics targeted for different populations (e.g., pregnant women with diabetes where the recommended TIR is 63 - 140 mg / dl). In aspects of embodiments of the present invention, these time-in ranges are used as input features for the proposed clustering algorithm.

[0095] The time-in range of a single daily CGM profile was used as the input for a single daily CGM profile when performing hierarchical clustering. The input was generated using all daily CGM profiles in the training dataset. The implementation of hierarchical clustering by the scipy cluster hierarchy Python module

[35] centroid algorithm was used to calculate the Euclidean distance between two rows of the input. Since the inventors wanted the hypoglycemic region behavior to be faithfully captured by the CSC, the inventors weighted the T54 and T70 input columns more than the TIR, T180, and T250 input columns. Each clinical similarity cluster (CSC) is a collection of daily CGM profiles such that each daily CGM profile in the collection basically has the same time-in range.

[0096] Figure 4 shows an exemplary process used to identify and then evaluate a single set of candidates for the CSC. For a given set of inputs (determined using the daily CGM profiles of the training data set and the weights selected for columns T54 and T70), a hierarchical clustering algorithm can generate a dendrogram that indicates the hierarchical relationships among the daily CGM profiles of the training data set. By "cutting" the dendrogram at a particular height, clustering by a particular number of clusters can be generated. The evaluation of each candidate set for the CSC can use the validation data set. The centroid of the CSC was used to classify each daily GCM profile in the validation data set. TIFF2025516193000030.tif6150 for an individual from TIFF2025516193000031.tif6150 Time-in range for the 6150th daily CGM profile for TIFF2025516193000032.tif as TIFF2025516193000033.tif6150, and TIFF2025516193000034.tif6150 for an individual from TIFF2025516193000035.tif6150 Time-in range for the CSC into which the 6150th daily GCM profile is classified for TIFF2025516193000036.tif as TIFF2025516193000037.tif6150, where it is TIFF2025516193000038.tif6150. Furthermore, as TIFF2025516193000039.tif12150, and assuming TIFF2025516193000040.tif12150, TIFF2025516193000041.tif6150 is for an individual The set of all daily CGM profiles of TIFF2025516193000042.tif6150. The inventors sought a single "optimal" set of CSCs. 1. For each set of points for TIFF2025516193000043.tif6150 maximize the r-squared value of linear regression through TIFF2025516193000044.tif6150, 2. Make the absolute value of the relative effect size less than 0.15 for each TIFF2025516193000045.tif6150, where the relative effect size is calculated as the average of the mean differences between TIFF2025516193000046.tif6150 divided by the standard deviation across all of TIFF2025516193000047.tif6150 and TIFF2025516193000048.tif6150.

[0097] The inventors explored nine different weightings of the T54 and T70 input columns (i.e., nine different sets of inputs to the hierarchical clustering algorithm), resulting in 166 candidate sets of CSCs to be evaluated. The final fixed set of CSCs has 35 clusters. Each CSC is defined by its centroid. The centroid of a given CSC is calculated using the daily CGM profiles in the training data set assigned to that CSC. The centroid of each CSC can be visualized as a CGM-based target.

[0098] Figure 5 shows the CGM-based target visualizations associated with each of the centroids of the 35 CSCs. The CSC centroid visualizations in Figure 5 are ordered by these TIR values (highest on the left and lowest on the right). Examination of this figure reveals that, as required, no two CSC visualizations are the same. Figure 6 shows the set of points for each TIFF2025516193000049.tif6150 The scatter plot of TIFF2025516193000050.tif6150 is shown, and Table 2 shows that when 35CSC is used for the set of 141,869 daily CGM profiles from the test dataset excluding the 1,169 daily CGM profiles in the test dataset belonging to healthy individuals, each provides the relative effect size for TIFF2025516193000051.tif6150. These results indicate that the final selected set of 35CSC faithfully represents the clinical characteristics of the daily CGM profiles to which they are similar. Table 2: Relative effect size and linear regression results for TIFF2025516193000052.tif6150 when using daily CGM profiles from the test dataset (excluding daily CGM profiles from healthy individuals) TIFF2025516193000052.tif6150 TIFF2025516193000053.tif63153

[0099] Table 3 of the 2019 International Consensus Time in Range

[14] defines the guidance for two diabetes groups. 1. Adults with T1D or T2D, and 2. Elderly / high-risk individuals with T1D or T2D Table 3: Fitted values of the parameters γ, λ, and k in the modified Weibull equation for six different sub-loops TIFF2025516193000054.tif57153

[0100] CSC1 meets the guidance for adults with T1D or T2D, and CSC5 meets the guidance for elderly / high-risk individuals with T1D or T2D. Therefore, physicians have target CSCs for these two situations.

[0101] Tracking of blood glucose variability Individual CSCs visited over time Each individual TIFF2025516193000055.tif6150 has a single CGM time series generated during a study in which each individual participated. Note that the CGM time series can have periods during which CGM data is not collected (e.g., during the washout period of the study). By classifying all of the daily CGM profiles into an individual's single time series, a sequence of indices (which may not be continuous if possible) that indicate the CSC to which each daily CGM profile is classified results in TIFF2025516193000056.tif16153. Each entry in the sequence TIFF2025516193000057.tif16153 corresponds to a single day of observation, and the days are ordered by the date of occurrence. Let TIFF2025516193000058.tif16153 be the number of unique CSCs visited by an individual t days after the observation by TIFF2025516193000059.tif16153. Next let TIFF2025516193000060.tif16153 be the trace of the number of unique CSCs visited by an individual over time by TIFF2025516193000061.tif16153, which provides an ideal of the variability in an individual's blood glucose. In general, a large number of unique CSCs visited by an individual indicates an increase in blood glucose variability (and thus poor blood glucose control). Note that in extreme cases, an individual with large blood glucose variability may visit only a small number of CSCs, and the visited CSCs may represent large blood glucose variability.

[0102] Average trace The average trace indicates the average behavior of individuals in a subgroup and can help emphasize differences in behavior between different subgroups. For a given subgroup for TIFF2025516193000062.tif16153, the average value is given by the following equation. TIFF2025516193000063.tif27153 where TIFF2025516193000064.tif16153 is the total number of individuals in the subgroup. All individuals in the subgroup do not necessarily have the same number of days observed in CSC index's own sequence TIFF2025516193000065.tif16153. Therefore, TIFF2025516193000066.tif16153 is defined only when there is a minimum number of sequences of time TIFF2025516193000067.tif16153. The minimum number of sequences is a function of the subgroup TIFF2025516193000068.tif16153.

[0103] Generally, the inventors want to estimate the long - term (more than 2 - 3 months) behavior of subgroups, especially subgroups without long - term data. To achieve this, the inventors fit a curve to the average trace of the subgroup. As seen in the top three rows of the plot in Figure 7, the curve has the general shape of a cumulative density function, and the inventors thought that a Weibull cumulative distribution function (CDF)

[36] modified to account for a non - unity upper limit value might provide a good fit. The following modified Weibull CDF was used as input to the curve_fit function of the scipy optimization SciPy module. TIFF2025516193000069.tif17153

[0104] Note that TIFF2025516193000070.tif16153 corrects the upper limit value so that 1 disappears from the Weibull CDF. TIFF2025516193000070.tif16153.

[0105] Figure 7 shows individuals, averages, and fit curves for each health state and treatment modality. The gray dashed - line curves in the first three rows of the plot in Figure 7 are individual traces TIFF2025516193000071.tif16153, and the thick solid line in the first three rows of the plot in Figure 7 is the average trace TIFF2025516193000072.tif is 16153. The thick solid line in the plot of the lower row of Figure 7 is the average trace TIFF2025516193000073.tif is 16153, and the thick dashed line is the modified Weibull curve fitted to each average trace (the parameters of these fitted curves can be found in Table 3). Since these curves track the rate of change of blood glucose, they distinguish between both the state of health and the treatment modality once again. As expected, the average healthy individual has the lowest blood glucose variability, followed by individuals with T2D, and then individuals with T1D. Among individuals with T1D, on average, individuals using MDI as a treatment modality have the highest blood glucose variability, followed by individuals undergoing PMP treatment, and then individuals undergoing CLC treatment.

[0106] The fitted modified Weibull curves reach a threshold after approximately 200 days of observation, and the maximum number of unique CSCs visited is higher in subjects with T1D than in those with T2D (13.3 CSCs vs. 9.2 CSCs) and in subjects treated with MDI vs. PMP or CLC (14.4 CSCs vs. 13.5 CSCs vs. 8.3 CSCs respectively).

[0107] Applications Data construction and dimensionality reduction All possible daily CGM profiles clinically represented by the AGP and this time-in-range are reduced to a finite fixed set of CSCs that can be used as inputs to decision support, clinical, and automated treatment algorithms.

[0108] Database indexing A database indexed by the structure defined by the CSCs can guarantee a fast and efficient search of subgroups of similar daily CGM profiles. This enables the realization of new features in decision support or automated insulin delivery systems, such as algorithm learning from a person's CGM patterns and from the patterns of other patients stored in a population database.

[0109] Daily CGM Profile Compression / Encryption Instead of transmitting the daily CGM profile (typically 288 data points), a single number can be transmitted that identifies the CSC index of the original daily CGM profile. At the receiving end, a decoder equipped with a set of CSCs can reconstruct the time-in-range and other clinical metrics of the daily CGM profile with a fidelity that preserves these metrics of the original daily CGM profile, such as AGP clinical characteristics.

[0110] Differentiating Health Status and Treatment Regimen One application of CSCs is the ability to differentiate health status and treatment regimen. Figure 8 shows the frequency distribution of 143,036 daily CGM profiles in a test data set across 35 different CSCs. The results of these plots are stratified by health status and treatment regimen. It is clear that the frequency distribution is a function of the health status of the individual being considered (i.e., a healthy individual, or an individual with T1D or T2D). In addition, for individuals with T1D, the frequency distribution is a function of the treatment regimen (MDI, PMP, or CLC). As expected, the majority (94.6%) of the daily CGM profiles generated by healthy individuals are classified as CSC1, and more than 99% of the daily CGM profiles are classified as one of CSC1, 2, or 3. The comparison of health status is made between T1D-MDI and T2D-MDI so that the comparison is fair, and since the treatment regimens are the same, the difference in the frequency distribution should be due to the difference in diabetes type (T1D vs T2D).

[0111] To formally test visual observations, an independent-sample Kruskal–Wallis test was performed between the frequency distributions of T1D-MDI, T1D-PMP, T1D-CLC, T2D-MDI, and the healthy subgroup, yielding p <.001. Pairwise comparisons with Bonferroni correction adjustment for multiple tests revealed adjusted significance of p < 0.05 for all pairwise comparisons except between T1D-PMP and T2D-MDI (see Figure 9), indicating that pump therapy results in clinical treatment outcomes in T1D-MDI patients similar to those in T2D, while at the same time showing that the clinical outcomes in T1D-CLC are superior to those in T2D.

[0112] Replacement of CGM for common clinical tests Measurement of fasting glucose levels, homeostatic model assessment (HOMA) of insulin sensitivity and β-cell function, or oral glucose tolerance test (OGTT) are common clinical methods for evaluating a person's glycemic health status. These, and other common glycemic function tests, generally require a doctor's visit, blood sampling, and laboratory analysis. In the case of OGTT, a test lasting several hours is required in a clinical setting. While being cumbersome, these tests are regularly required in many situations, for example, frequent OGTTs are required in gestational diabetes. The ability of CSC to distinguish with high fidelity between health states can function as an alternative to these clinical tests, and wearing a CGM for 10 days in a home environment, achieved by a pre-determined schedule of meals and physical activity, greatly simplifies data collection while achieving diagnostic results similar to those acceptable in clinical practice.

[0113] CGM pattern recognition and prediction The transition probability matrix representing the evolution of patients across a predefined CSC is a natural tool for observing the progression of disease or treatment. Pattern recognition, or recurrent behavior, is reflected by patterns or cycles detected in the transition probabilities from one state to the next. Short-term or long-term prediction of blood glucose control is based on probabilistic patterns or repeated visits to a certain subspace of the Markov chain state space. The latter is the subject of the theory of semi-Markov chains and results from the aggregation (ramping) of the state space into related subsets, characterized by the random durations consumed in each subset.

[0114] Tracking the progression of a disease over time Deterioration of the disease is indicated by a transition to an undesirable state, and conversely, successful treatment optimization by titration of a drug is reflected by a transition to a clinically desirable state. In actual applications, the state space of the Markov chain is defined / aggregated to correspond to the CSC defined by aspects of the embodiments of the systems, methods, and computer-readable media of the present invention.

[0115] Example 2 The ability to track blood glucose has evolved from intermittent self-monitoring to modern continuous glucose monitoring (CGM). However, clinical decisions based on CGM are difficult because CGM provides rich and complex datasets that require sophisticated analysis to provide insights. This work presents a two-step iterative process that reduces the “clinical dimensionality” of CGM data space by identifying a fixed set of clinical similarity clusters (CSCs), such that the daily CGM profiles within each cluster convey similar clinical messages. The two-step process uses hierarchical clustering on a training dataset to identify a candidate set of CSCs, evaluates the ability of the candidate set of CSCs using linear regression and relative effect sizes, and incorporates five different time-in-ranges of daily CGM profiles from a validation dataset. The optimal set of 35 CSCs identified using the validation dataset was then used to classify daily CGM profiles in another test dataset. The results show that the set of CSCs is robust, generalizes well, and most importantly captures the clinical characteristics of daily CGM profiles with high fidelity. This fixed set of CSCs enables tracking of an individual's daily blood glucose control over time, facilitates the design of individualized treatment, and potentially enables automated treatment optimization according to predefined rules that map optimal treatment responses to each CSC. CSCs can also be used to visualize differences in blood glucose control between individuals and differences between treatment modalities, and to identify individuals who can benefit from treatment adjustments.

[0116] Glucose variability (GV) in diabetes reflects the biological behavioral processes that cause blood glucose (BG) variability, which has two major dimensions: amplitude, which represents the degree of BG excursions, and time, which represents the frequency of BG fluctuations and the rate of event progression. Observation of this process has evolved from sporadic self-monitoring that generates several BG readings per day to the latest continuous glucose monitoring (CGM) that generates large datasets, time-series of glucose readings at regular intervals (e.g., every 5 minutes). As the diffusion of CGM technology increases, inevitably a vast amount of data is generated. CGM time-series data are used to gain insights that enable appropriate treatment of diabetes, including risk stratification, prediction of events of interest (e.g., impending hypoglycemia or hyperglycemia), or what is commonly referred to as "artificial pancreas", an automated closed-loop control. To improve the clinical use of CGM data and simplify these interpretations, the 2019 Time in Range (TIR) International Consensus proposed TIR as the primary CGM-based measure of glycemic control, set clinical goals for its use, and this "TIR measure system" has been widely adopted over the past three years. The TIR system is based on the ambulatory glucose profile (AGP), which was introduced as a template for data presentation and visualization. Originally proposed by Mazze et al. for sporadic self-monitoring data, the standardized CGM report incorporates the core CGM measures and goals, along with a 14-day composite glucose profile, as essential components for clinical decision-making.

[0117] One area of existing literature focuses on using CGM data to cluster subjects (e.g., clustering subjects into groups at high risk of gestational diabetes) or to build classification models (e.g., classifying subjects as healthy, pre-diabetic, or diabetic based on their own CGM data). Acciaroli et al. used 25 CGM-based glycemic variabilities as inputs to a two-step binary logistic regression model. The model first classifies subjects as healthy or not healthy, and then classifies subjects not classified as healthy in the first step as being affected by either impaired glucose tolerance (IGT) or type 2 diabetes (T2D). The model was able to distinguish between people with health and IGT or T2D, and also between IGT and T2D. Bartolome et al. developed an algorithm called Glucamine for the purpose of discovering individualized patterns in long-term CGM data (3 - 6 months of data) that do not appear in short-term data. Gecili et al. used functional data analysis to identify phenotypes of glycemic variability in type 1 diabetes (T1D) using CGM data. They concluded that these phenotypes can be used to optimize T1D management for subgroups of subjects at high risk of adverse outcomes. Inayama et al. derived summary statistics from GCM data and then used hierarchical clustering of the summary statistics to cluster 29 women into three groups (low hypoglycemic levels with low glycemic variability, L; medium glycemic levels with medium to high glycemic variability, M; high glycemic levels with high glycemic variability, H). They found that women with gestational diabetes (GDM) tend to fall into the H group, and thus their clustering can help identify subgroups of women with characteristics of GDM. Liet et al. took CGM time series data and decomposed it into trend, season (daily), and random components. The trend component was then clustered using K-means clustering, resulting in five clusters, two of which had an increasing trend, two of which had a decreasing trend, and one of which had no trend (no change).Next, they used these five clusters to group the subjects into three groups: an increasing group, a decreasing group, and a non-changing group. The fasting blood glucose levels of the subjects assigned to the increasing and decreasing groups increased and decreased, respectively, following a six-month long-term glycemic reduction treatment. Tao et al. clustered the 24-hour CGM time series generated by T2D subjects for the purpose of identifying subjects with dysglycemia and various degrees of clinical phenotypes. Mao et al. developed a pipeline for the analysis of CGM data for the purpose of identifying glucotypes, such as groups of subjects that differ in their degree of control, the amount of time spent in range, and the presence and timing of hypoglycemia and hyperglycemia. They state that their method can be used "to introduce targeted interventions among patients with diabetes" in addition to other biometric data.

[0118] Other literature uses CGM data to construct analytical frameworks that can be used for several different applications (e.g., identifying different subtypes of patients). Hall et al. developed an "analytical framework" that can group individuals according to specific patterns of glucose response called "glucotypes" that reveal heterogeneous, or secondary phenotypes, within traditional diagnostic categories of glucose regulation. Matabeuna et al. used CGM data to derive glucodensity, which they claim is the distribution of blood glucose values from the CGM time series of individual subjects, where glucodensity is an extension of the time-in-range metric. They propose that glucodensity can be used in clinical practice to provide "an accurate representation of an individual's glucose profile," "identify different subtypes of patients based on their glycemic state and other variables," and "ascertain whether there are statistically significant differences among patients receiving various interventions."

[0119] All of the documents cited above have the ultimate goal of classifying individuals with diabetes into distinct subgroups, and none of the classifications incorporate the daily variations in blood glucose control within an individual. To more comprehensively utilize the temporal structure of CGM data, the inventors define a daily CGM profile, a 24-hour CGM time series, and use these to establish a finite set of 483 representative daily profiles or motifs, which they claim can match almost any daily CGM profile. The set of motifs is externally validated and can be considered fixed. The motifs reflect differences not only between individuals but also in the daily variations in an individual's blood glucose control.

[0120] The work proposed in this document constructs an analytical framework based on a "TIR scale system" that enables the classification of an individual's daily blood glucose behavior. This classification for a single day provides the basis for a number of different analyses, which can be used to define a fixed number of groups if needed, but can also be used, for example, for modeling purposes, clinical subgroup stratification and transitions from one subgroup to another, or to track an individual over time for the notification of an automated control method. Additionally, existing literature assumes that the subject has a particular type of diabetes, whereas the analytical framework presented in this document is agnostic to the type of diabetes and works equally well with CGM data generated by healthy individuals, and the patterns observed can identify an individual's blood glucose state without the need for prior classification or diagnosis.

[0121] The data and methods used to identify clinically similar clusters (CSCs) are outlined herein, including the use of training and validation datasets to obtain a fixed set of CSCs and then evaluating the performance of the fixed set of clusters in another test set.

[0122] Method The method includes steps of defining input generation for a hierarchical clustering method used to identify training, validation, and test data sets; a candidate set of clinical similarity clusters (CSCs); and a two-step iterative process used to identify an “optimal” set of CSCs.

[0123] Data set source The data used in this work was generated from the following. 1. University of Virginia Diabetes Technology Center (i.e., DCLP1[4], DCLP3[5], DIA1

[21] , DIA2

[22] , DSS1

[23] , NTLT

[24] , and TRLNT studies), and 2. Health Research Jaeb Center website (i.e., CITY

[25] , DCLP5[6], NDIAB

[26] , MDEX

[27] , REPBG

[28] , RTCGM[29, 30], SENCE

[31] , SEVHYPO

[32] , and WISDM[33, 34] studies). For these studies, the analysis, content, and conclusions presented in this work are solely the responsibility of the authors and have not been reviewed by any of the study groups.

[0124] From each of these studies, the procedures outlined in the inventors' previous literature

[19] have been used to process the CGM time series and define the daily CGM profile, where the daily CGM profile is a time series of 288 blood glucose data points collected every 5 minutes during a midnight-to-midnight (24-hour) period. From 16 datasets, there were 2,462 subjects and a total of 204,710 daily CGM profiles. The studies represent healthy individuals, individuals with T1D, and individuals with T2D. The studies also represent a variety of treatment modalities including multiple daily injections (MDI), insulin pumps (CSII), and closed-loop control (CLC). The characteristics of the participants in each study are detailed in Table 4. Most of the daily GSM profiles (95.9%) were generated by subjects with T1D treated by either MDI, CSII, or CSC. Two datasets (DCLP5 and SENCE) focused on children with T1D. The Dia2 dataset included data from participants with T2D during MDI treatment and represented 3.6% of the generated daily CGM profiles. The NDIAB and TRLNT datasets included data from people without diabetes (healthy). The glycemic control, evaluated by the mean HbA1c of each participant at baseline, fell between 5.2% (NDIAB study) and 9.1% (City study) for participants without diabetes. Healthy people generally had data for less than 7 days, and people in the majority of the other studies had data for an average of 5 or more weeks.

[0125] The 204,710 daily CGM profiles were used to form three different datasets, the training dataset, each with a distinct purpose. The training dataset, this dataset was composed of 23,916 daily CGM profiles taken from the DCLP1, DCLP3, DIA1, DIA2, DSS1, and NTLT studies and was used to define a candidate set for CSC. Table 4: Characteristics of the 16 datasets used in this study. Statistical values are shown as mean (SD) unless otherwise indicated. T1D is type 1 diabetes, T2D is type 2 diabetes, BMI is body mass index, CGM is continuous glucose monitoring, MDI is multiple daily injections, CSII is insulin pump, and CLC is closed-loop control. * indicates that data was not available at the subject level and was thus taken from the study protocol. Study name Health status Number of study subjects Age, years BMI HbA1c Treatment mode Study duration Number of daily GCMs TIFF2025516193000074.tif2381701. Validation dataset: This dataset was composed of 37,758 daily CGM profiles taken again from the DCLP1, DCLP3, DIA1, DSS1, and NTLT studies and was used to a) evaluate the performance of each candidate set of CSCs and b) select the final and fixed set of CSCs. 2. Test dataset: This dataset was composed of 143,036 daily CGM profiles taken from the City, DCLP5, DIA2, MDEX, NDIAB, REPBG, RTCGM, SENCE, SEVHYPO, TRLNT, and WISDM studies and was used to evaluate the robustness and generalizability of the final selected set of CSCs.

[0126] Hierarchical clustering to identify clinically similar clusters of daily CGM profiles. The TIR scale system defines five time-in-ranges of blood glucose values. That is, 1. Level 2 hypoglycemia (T54): Blood glucose strictly less than 54 mg / dl, 2. Level 1 hypoglycemia (T70): Blood glucose higher than or equal to 54 mg / dL and strictly less than 70 mg / dL, 3. Target range (TIR): Blood glucose higher than or equal to 70 mg / dL and less than or equal to 180 mg / dL, 4. Level 1 hyperglycemia (T180): Blood glucose strictly higher than 180 mg / dL and less than or equal to 250 mg / dL, 5. Level 2 hyperglycemia (T250): Blood glucose strictly higher than 250 mg / dL.

[0127] The five time-in-ranges of each daily CGM profile were used as input for hierarchical clustering and were computer-calculated by the scipy.cluster.hierarchy Python module

[35] , where the centroid method was used to calculate the Euclidean distance between two rows of the input. The T54 and T70 input columns were weighted multiplied by TIFF2025516193000075.tif16153 and TIFF2025516193000076.tif16153 to emphasize the importance of these two input columns during the clustering process and to ensure that hypoglycemic-range behavior (i.e., T70 and T54) was captured with high fidelity. Each clinically similar cluster is a collection of daily CGM profiles such that each daily CGM profile in the collection has essentially the same time-in-range. The centroid of each cluster of daily CGM profiles identified by hierarchical clustering can define the CSC. In contrast to the inventors' previously published work ([19,20] see) using daily GCM profiles, the CSC ignores the timing within a day of blood glucose fluctuations. Thus, reducing a daily CGM profile composed of 288 data points to a single CSC centroid composed of only 5 data points, one for each of the 5 time-in-ranges, involves extracting the timing information contained within the daily CGM profile.

[0128] An iterative procedure to identify the "optimal" set of CSCs FIG. 10 shows an exemplary two-step iterative process used to identify the “optimal” set of CSCs. The input used to define the candidate set of CSCs was generated using 23,916 daily CGM profiles in the training dataset. For a given set of inputs (determined using the daily CGM profiles in the training dataset and the weights selected in the T70 and T54 columns), the hierarchical clustering algorithm can generate a dendrogram indicating the hierarchical relationships between the time ranges of the daily CGM profiles in the training dataset. By “cutting” the dendrogram at a specific height, a specific set of N clusters can be defined, and the centroid of each cluster is calculated using the daily CGM profiles assigned to that cluster. This set of N clusters is then the candidate set of CSCs that must be evaluated next.

[0129] The evaluation of each candidate set of CSCs begins by classifying the 37,758 daily CGM profiles in the validation dataset using the CSC centroid. TIFF2025516193000077.tif6150 for an individual From TIFF2025516193000078.tif15153 The time range of the 17153rd daily CGM profile of TIFF2025516193000079.tif As TIFF2025516193000080.tif16153, and TIFF2025516193000081.tif17153 for an individual From TIFF2025516193000082.tif6150 The time range of the CSC centroid to which the 6150th daily CGM profile of TIFF2025516193000083.tif is classified As TIFF2025516193000084.tif16153, where TIFF2025516193000085.tif16153. Furthermore, Set it as TIFF2025516193000086.tif26153, If we set it as TIFF2025516193000087.tif22153, then here TIFF2025516193000088.tif16153 is for individuals TIFF2025516193000089.tif16153 is a set of all daily CGM profiles of an individual. Given the above definitions, the "optimal" set of CSCs is as follows. 1. For each TIFF2025516193000090.tif16153, TIFF2025516193000091.tif17153 having a relative effect size less than or equal to this, 2. For each TIFF2025516193000092.tif11150, a set of points TIFF2025516193000093.tif16153 maximizes the linear regression TIFF2025516193000094.tif17153 value through the intercept.

[0130] The relative effect size is computer-calculated as follows. TIFF2025516193000095.tif26153 Here, the denominator is the standard deviation of all 37,758 TIFF2025516193000096.tif17153 values. Linear regression by intercept (specifically, the scipy.stats.linregress function from the SCiPy

[35] Python package) was used during this part of the process to TIFF2025516193000097.tif6150 compare values. Table 5: When the 35 CSCs of TIFF2025516193000098.tif16153 are used to classify the 37,758 daily CGM profiles (upper) in the validation dataset and the 143,036 daily CGM profiles (lower) in the test dataset, the relative effect sizes and linear regression results of five different time-in-ranges {T54, T70, TIR, T180, T250} are presented. Note that the presented linear regression results are linear regressions with a fixed intercept of 0. TIFF2025516193000099.tif82153

[0131] In this application, the relative effect size TIFF2025516193000100.tif16153 is an attempt to ensure that a) relatively rare hypoglycemia events are captured with reasonably high fidelity, i.e., the swap of daily CGM profiles by the CSC results in a deviation of no more than 15%, and b) the set of CSCs captures the blood glucose dynamics regardless of the individual generating the daily CGM profile.

[0132] Results An “optimal” set of clinical similarity clusters is identified, and the performance of this set in the test dataset of daily CGM profiles is also presented.

[0133] Identification of a set of clinical similarity clusters Nine different combinations of weights TIFF2025516193000101.tif17153 were explored. Upper bounds of 60 and lower bounds of 15 were used to limit the number of candidate sets of CSCs considered: Any “cuts” of the hierarchical clustering dendrogram that resulted in candidate sets of CSCs with more than 60 or fewer than 15 CSCs were not evaluated using the second step of the iterative procedure. There were 166 candidate sets of CSCs that were evaluated, and the final selected (fixed) set of CSCs, TIFF2025516193000102.tif16153 has 35 clusters. The centroid of a given CSC is calculated using the daily CGM profiles in the training dataset assigned to this CSC. Using the 37,758 daily CGM profiles of the validation dataset The relative effect sizes and linear regression results from evaluating TIFF2025516193000103.tif16153 are shown in the upper half of Table 5. The linear regression results are for linear regression with a fixed intercept of 0 and are achieved using the sklearn.linear model, the linear regression function of the scikit-learn Python package. When TIFF2025516193000104.tif16153 was used to classify the 37,758 daily CGM profiles in the validation dataset, each Set of points for TIFF2025516193000105.tif16153 Scatter plot of TIFF2025516193000106.tif16153. These results TIFF2025516193000107.tif17153, the final selected set of 35 CSCs, shows that it faithfully represents the clinical characteristics of the approximated daily CGM profiles. In particular, the CSCs capture the relatively rare hypoglycemia component of the daily GCM profiles with high fidelity.

[0134] The centroid of the CSC can be visualized as a CGM-based target and is replicated in the legend in Figure 11. The altitude of each color in the visualization corresponds to the percentage of time the centroid spends in each range. Figure 11 is ordered by its TIR value TIFF2025516193000108.tif shows the CGM-based target visualization associated with each center of gravity of the 35 CSC centers of gravity at 16153 (highest on the left and lowest on the right). Examination of this figure reveals that, as required, the two CSC centers of gravity are not the same, indicating that the two CSCs are not the same. In addition, further examination reveals that the majority of the centers of gravity have significant hypoglycemia (T54 and T70) for them. This is the result of the weighting of the T54 and T70 input columns when generating the candidate set of CSCs in the first step of the iterative procedure.

[0135] Next The centers of gravity of the 35 CSCs at 17153 in TIFF2025516193000109.tif were used to classify the 143,036 daily CGM profiles in the test dataset. Figure 12 shows each Set of points for TIFF2025516193000110.tif at 16153 TIFF2025516193000111.tif illustrates 16153, and the lower half of Table 5 provides the associated linear regression results and relative effect sizes. These results TIFF2025516193000112.tif shows that 17153 is robust, has not been seen before, and generalizes to data that has not been trained in the case of healthy individuals.

[0136] Visualization of an individual's blood glucose control CSCs can be used to visualize differences in blood glucose control among individuals with the same health status and treatment modality, and thus can identify individuals who may require more individualized attention. Let the formula be TIFF2025516193000113.tif be 21153, where TIFF2025516193000114.tif at 17153 is the individual Of TIFF2025516193000115.tif at 17153 TIFF2025516193000116.tif16153 is the number of unique CSCs required to classify the CGM profiles by day. Since the total number of unique CSCs is limited (i.e., there are only 35 different CSCs), if TIFF2025516193000117.tif16153 is large, then TIFF2025516193000118.tif16153 tends to be zero. To overcome this problem, TIFF2025516193000119.tif16153 was fixed to the 28-day profile (i.e., four weeks of data). Next if TIFF2025516193000120.tif17153 TIFF2025516193000121.tif16153 = the average value of the values generated by computer calculating each TIFF2025516193000122.tif16153 using a sliding window of the 28-day CGM profile, where TIFF2025516193000123.tif16153 advances every 7 days (one week of data), and each sliding window TIFF2025516193000124.tif16153 must computer calculate at least the 14-day CGM profile (two weeks of data) of TIFF2025516193000125.tif16153. For example, if an individual TIFF2025516193000126.tif16153 has a 40-day CGM profile, TIFF2025516193000127.tif16153, TIFF2025516193000128.tif16153, TIFF2025516193000129.tif16153, and TIFF2025516193000130.tif16153 becomes the average value, where TIFF2025516193000131.tif16153 is computer-calculated using daily CGM profiles 1 through 28, TIFF2025516193000132.tif16153 is computer-calculated using daily CGM profiles 8 through 35, TIFF2025516193000133.tif16153 is computer-calculated using daily CGM profiles 15 through 40, TIFF2025516193000134.tif16153 is computer-calculated using daily CGM profiles 22 through 40. Note that when the number of daily CGM profiles in the sliding window is less than that of TIFF2025516193000135.tif16153 (it has only 18 daily CGM profiles like that of TIFF2025516193000136.tif17153's sliding window), the boundary of equation (4) is then set to the number of daily CGM profiles in the sliding window. Note that different CSCs are distinct (i.e., the two CSC centroids do not have the same time range by configuration), so TIFF2025516193000137.tif16153 is a measure of how unstable an individual's daily blood glucose control is. 1 / close to TIFF2025516193000138.tif16153 The value of TIFF2025516193000139.tif16153 indicates that the individual has not visited a large number of different CSCs, and as a result, the individual does not have a high variability in their daily blood glucose control, 1 / greater than TIFF2025516193000140.tif16153 The value of TIFF2025516193000141.tif16153 indicates that the individual has a high variability in their daily blood glucose control.

[0137] TIFF2025516193000142.tif16153 uses the sliding window of the daily CGM profile for TIFF2025516193000143.tif16153 for an individual TIFF2025516193000144.tif16153 of TIFF2025516193000145.tif16153 as the average CSC index of the daily CGM profile, TIFF2025516193000146.tif16153 for each TIFF2025516193000147.tif16153 is the average value of the values generated by computer calculation TIFF2025516193000148.tif16153. Here, the sliding window advances every 7 days, and further, each sliding window TIFF2025516193000149.tif16153 must computer-calculate at least 14 daily CGM profiles. Since the CSCs are indexed such that the centroid of CSC1 has the highest TIR value and the centroid of CSC35 has the lowest TIR value, TIFF2025516193000150.tif17153 is a measure of the individual's TIFF2025516193000151.tif16153 daily continuous glucose control. A value close to 1 TIFF2025516193000152.tif16153 indicates that the individual spends more time in the target range, and a value close to 35 indicates that the individual spends less time in the target range.

[0138] Figure 12 is for classifying 141,867 daily CGM profiles of the test dataset TIFF2025516193000153.tif16153 resulting from using TIFF2025516193000154.tif16153 points for TIFF2025516193000155.tif shows an exemplary scatter plot of 16153.

[0139] Figure 13 shows a 2D histogram of hex bins, where the bins are hexagonal and the color is the point TIFF2025516193000156.tif represents the number of data points within each bin for 17153. For individuals with T1D or T2D, the point TIFF2025516193000157.tif 16153 is TIFF2025516193000158.tif 16153, i.e., the individual TIFF2025516193000159.tif 16153 was generated only if the individual had at least 14 daily CGM profiles. This constraint is generally for healthy individuals since these individuals do not generally have more than 7 daily CGM profiles TIFF2025516193000160.tif 16153 is relaxed for (see NDIAB and TRLNT rows) - in this case, TIFF2025516193000161.tif 16153 and TIFF2025516193000162.tif 17153 were computer calculated using only a single sliding window.

[0140] The upper left figure in Figure 13 is the point for all individuals TIFF2025516193000163.tif 16153 is shown, and the minimum value shows a wide range of blood glucose control shown by the individual. The points of individuals with different combinations of health states and treatment modalities Other plots illustrating TIFF2025516193000164.tif 16153 of this point also show that there is a wide variety of blood glucose control even when a single combination of health state and treatment modality is considered. As expected, the points of healthy individuals TIFF2025516193000165.tif 16153 are positioned towards the lower left hand side of the plot. Large The value of TIFF2025516193000166.tif16153 is due to a small number of days of CGM data applicable to healthy individuals. Generally, the inventors expect that the point of TIFF2025516193000167.tif16153 will be positioned towards the bottom left corner of the plot. The differences in the plots, as expected, generally imply that individuals with T1D during CLC therapy have better glycemic control than individuals with T1D during CSII therapy, and individuals with T1D during CSII treatment have better glycemic control than individuals with T1D during MDI therapy.

[0141] TIFF2025516193000168.tif16153 is small, but if TIFF2025516193000169.tif16153 is large, then note that the individual does not visit a large number of different CSCs and does not spend much time in the target range. This is the worst-case scenario and is represented by the individual in the upper left corner of the "T1D individual" plot. Further analysis shows that this individual spends 89% of the 74-day profile on CSCs with an index greater than or equal to 30, and 53% of the 74-day CGM profile is classified as CSC35. This individual is a prime example of someone who can benefit from more individualized clinical attention.

[0142] In this work, the inventors present a two-step iterative process for identifying a fixed set of clinically similar clusters (CSCs) of daily CGM profiles. The two-step process uses hierarchical clustering on a training dataset for identifying a candidate set of CSCs and a validation dataset for evaluating the performance of the candidate set of CSCs. In particular, the ability of the CSCs to faithfully capture five different time-in-ranges of the daily CGM profiles to be classified is evaluated. Then a fixed set of 35 CSCs TIFF2025516193000170.tif16153 was used to classify daily CGM profiles in another test dataset, and the result that the set is robust and generalizes properly was shown. Additionally, the distribution of different CSCs of daily CGM profiles is shown to be unique to health status and treatment modalities.

[0143] There are potential applications for a wide variety of CSCs. All daily CGM profiles, clinically represented by their own time ranges, are reduced to a finite and fixed set of CSCs that can be used as inputs to decision support, clinical, and automated treatment algorithms. Furthermore, a database indexed by the structure defined by the CSCs guarantees a fast and efficient search of subgroups of clinically similar daily CGM profiles. CSCs also have the potential to enable new features in decision support, to track changes in an individual's daily blood glucose control, or to enable notifications for automated insulin delivery systems such as algorithm learning from a person's CGM patterns and from the patterns of other patients stored in a population database. Finally, abstracting the typical 288 data points of a daily CGM profile to a single CSC index has aspects of both data compression and encryption. These applications are the subject of further research.

[0144] The following eight datasets were downloaded from https: / / public.jaeb.org / datasets / diabetes: CITY, DCLP5, MDEX, REPBG, RTCGM, SENCE, SEVHYPO, WISDM. The analysis, content, and conclusions presented in this work are solely the responsibility of the authors and have not been reviewed by any of the research groups (CITY: CGM Intervention Research Group in Adolescents and Young Adults with T1D; DCLP5: iDCL Trial Research Group; MDEX: T1D Exchange Mini-Dose Glucagon Exercise Research Group; REPBG: REPLACEBG Research Group; RTCGM: JDRF CGM Research Group; Sence: Research Group on Strategies to Expand the Use of New CGM in Infancy; SEVHYPO: T1D Exchange Severe Hypoglycemia Research Group in Older Adults with Type 1 Diabetes; WISDM: WISDM Research Group). Table 6: T54 and T70 Weighting Combinations - Weighting TIFF2025516193000171.tif17153 and Nine different combinations of TIFF2025516193000172.tif16153 were used to highlight the T54 and T70 input columns when performing hierarchical clustering. TIFF2025516193000173.tif70153

[0145] Example 3 Background: The adoption of CGM results in a vast amount of data, but the interpretation of these is more of an art than a science itself. The International Consensus on Time in Range (TIR) has proposed a widely accepted TIR scale system, and the inventors here advance by introducing a finite and fixed set of clinically similar clusters (CSCs), whereby the TIR scale of the daily CGM profiles within a cluster is uniform.

[0146] Methods: CSC definition and validation were performed using 204,710 daily CGM profiles in healthy, type 1 and type 2 diabetes (T1D, T2D) under different treatments. CSC was defined using 23,916 daily CGM profiles (training data), and the final fixed set of CSC was obtained using another 37,758 profiles (validation data). Test data (143,036 profiles) were used to establish the robustness and generalizability of CSC.

[0147] Results: The final set of CSC contains 35 clusters. Any daily CGM profile could be classified into a single CSC that faithfully approximated the common blood glucose measures of the daily CGM profile, as demonstrated by regression analysis with a 0 intercept (R-square > 0.81, e.g., correlation > 0.9 for all TIR and many other metrics). CSC enabled the tracking of daily changes in a person's blood glucose control over time, distinguishing CGM profiles in healthy, T1D, and T2D under different treatments.

[0148] Conclusions: Any daily CGM profile can be classified into one of only 35 pre-fixed CSCs, which can support applications such as aggregated data interpretation and algorithmic approaches to treatment, CGM replacement for clinical testing, database indexing, pattern recognition, and tracking of disease progression.

[0149] Introduction The widespread adoption of continuous glucose monitoring (CGM) technology has inevitably generated an enormous amount of data. For example, two recent reports on the real-life use of artificial pancreas systems were based on over 1.5 billion data points. Over the years, several glycemic control metrics have been introduced with the general aim of aggregating CGM data to convey meaningful clinical messages. Some existing metrics based on self-monitoring data, such as MAGE (mean amplitude of glucose excursions) and LBGI / HBGI (low and high BG indices), have been similarly adapted for CGM use. The adaptation of MAGE to CGM data follows the classic time-dependent structure of this metric, so in this case, CGM is used only as a source of amplitude assessment. The adaptations of LBGI and HBGI were applied to the differences between SMBG and CGM data. The mean of daily differences (MODD) was introduced as a measure of daytime variability, and continuous overlapping net glycemic action (CONGA) was presented as a composite index of the magnitude and timing of glucose (BG) fluctuations captured over various periods. The standard deviation of the rate of change of BG was used as a marker of the stability of the metabolic system over time, based on the premise that abnormal BG changes are signs of system instability. Arrays of standard deviations were introduced to reflect the glycemic variability contained within different clinically relevant periods of CGM data. The clinical interpretation of various CGM-based glycemic variability metrics involved the analysis of CGM data using several graphs, including the Poincaré plot of system stability and variability grid analysis (VGA) used to visualize the glycemic variability captured by CGM and the effectiveness of automated insulin delivery (AID), as discussed in eight early reviews of statistical methods available for the analysis of CGM data. Perspectives published in reviews in Biomedical Engineering 11 and Diabetes Care 12 evaluated methods of computing and visualizing glycemic variability in the context of its relationship to the risk of hypoglycemia. The glucose management indicator (GMI) was introduced as a CGM-based approximation of HbA1c measurement. The latest addition to the series of glycemic control metrics is the glycemic risk index (GRI), which attempts to balance the risks of hypoglycemia and hyperglycemia in a single metric, based on the collective opinion of several physicians.

[0150] As a result, the CGM field has become overburdened not only by vast datasets but also by numerous metrics used to evaluate various aspects of glycemic control. For reference, many, but not all, of the existing metrics are discussed in detail in a 2017 paper published in Nature Reviews Endocrinology. A common theme was that CGM-based metrics generally must include not only the amplitude of CGM readings but also some notion of the timing of CGM readings. This is because CGM data enables analytical methods that go beyond the scope of time series of glycemic observations that are equally spaced in time, traditional MAGE, LBGI / HBGI, MODD, CONGA, GRI, or any other amplitude-based metric that has been accepted over the years. For example, the latest algorithms enabling AID are possible only because of the timing information carried by the CGM data stream.

[0151] More recently, it has been shown that by sorting through the clutter of multiple glycemic markers, virtually all glycemic control metrics introduced over the years can be represented by just two "basic metrics" - hyperglycemia exposure and hypoglycemia risk, meaning that the quantitative expression of glycemic control is a rather simple two-dimensional structure. The 2019 Time in Range (TIR, generally 70 - 180 mg / dL) International Consensus proposed TIR as the first CGM-based metric for glycemic control and set clinical goals for its use. If the hypoglycemia range (TBR), TIR, and hyperglycemia range (TAR) always add up to 100 percent, the TIR metric system is an excellent representation of the two-dimensional structure of glycemic control, with TBR as a measure of the risk of hypoglycemia and TAR (or, equivalently, TIR) measuring hyperglycemia exposure. This system was introduced as a template for data presentation and visualization originally developed by Mazze et al. for self-monitoring data, the Ambulatory Glucose Profile (AGP). Based on the AGP, the standardized CGM report incorporates the core CGM metrics and goals along with the 14-day composite glucose profile as an essential component of clinical decision-making. This format has been approved by the TIR17 International Consensus and referenced by the American Diabetes Association 2019 Standards of Care 20 and the AACE Consensus on the Use of CGM21. The AGP report is currently adopted by many CGM device manufacturers in CGM companion software and has been proposed by the International Consensus as a standardized output in the evaluation of AID technology and for the presentation of results from clinical trials.

[0152] The widely adopted TIR metric system defines five time in ranges for CGM glucose values. These time in ranges are used to provide the numerical interpretation of the AGP, namely, level 2 hypoglycemia - below 54 mg / dl, level 1 hypoglycemia - 54 to 69 mg / dL, time in range (TIR) - 70 to 180 mg / dL, level 1 hyperglycemia - 180 to 250 mg / dL, and level 2 hyperglycemia - above 250 mg / dL.

[0153] These boundaries can be changed according to the consensus recommendations for different types of diabetes, but the concept is the same. A static visual and quantitative representation - a (typically 14-day) snapshot of CGM data is thus appropriately established by the AGP / TIR representation. However, the TIR scale system does not reflect the day-to-day variability of CGM data (except for the points of the AGP cloud), or the progression (improvement / deterioration) of glycemic control over time. If the main advantage of CGM is to measure the time series of blood glucose values and capture the evolving process of glycemic control, it becomes most important to equip the TIR system with a time component. Thus, this manuscript takes the next step in advancing the AGP / TIR concept by establishing a fixed and finite set of clinically similar clusters (CSCs), faithfully representing many all daily CGM profiles with a relatively small (N = 35) fixed CSCs, and enabling the tracking of daily blood glucose changes over time in a table lookup format.

[0154] Materials and Methods Data Sixteen unidentifiable record literature datasets were used in this work as detailed in Table 7, which includes demographic information and summary statistics of study participants, type of diabetes (T1D, T2D), or health, and treatment modalities, e.g., multiple daily injections (MDI), continuous subcutaneous insulin delivery (CSII) via insulin pump therapy, or automated insulin infusion therapy (AID). Table 7 Characteristics of participants in the 16 studies used in this work TIFF2025516193000174.tif239163 TIFF2025516193000175.tif250155 The statistical values of TIFF2025516193000176.tif58160 are shown as mean (SD) unless otherwise indicated. T1D represents type 1 diabetes, T2D represents type 2 diabetes, BMI represents body mass index, CGM represents continuous glucose monitoring, MDI represents multiple daily injections, CSII represents continuous subcutaneous insulin infusion therapy via an insulin pump, and AID represents automated insulin delivery therapy.

[0155] These data were collected in clinical trials conducted at the University of Virginia Diabetes Technology Center or are available in the public data repository of the Jaeb Health Research Center in Tampa, Florida. Citations for these studies are as follows: CITY, 24 DCLP1, 25 DCLP3, 26 DCLP5, 27 DIAMOND1, 28 DIAMOND2, 29 DSS1, 30 MDEX, 31 NDIAB, 32 NIGHTLIGHT, 33 REPLACE - BG, 34 RTCGM, 35 SENCE, 36 SEVHYPO, 37 UVA - TRIALNET, 38, and WISDOM.39. Overall, these datasets included CGM traces of 2,462 individuals (52.6% female) and 204,710 daily CGM profiles, or approximately 560 person - years of data for health, T1D, and T2D. Many of the daily CGM profiles (95.9%) were generated by individuals with T1D treated with MDI, CSII, or AID. Two studies focused on children with T1D (DCLP5 and SENCE). The DIAMOND2 study had participants with T2D on MDI treatment and represented 3.6% of the daily CGM profiles. The NDIAB and UVA - TRIALNET studies included data from people without diabetes. Glycemic control, as evaluated by baseline mean HbA1c, ranged from 5.2% (NDIAB study) to 9.1% (CITY study) for people without diabetes. In health, individuals generally had data for less than 7 days, and participants in the majority of diabetes studies had data for an average of 5 or more weeks.

[0156] Data preprocessing and separation for training, validation, and test data Previously published procedures were used to process CGM time series and define daily CGM profiles, where the daily CGM profile is a time series of 288 blood glucose data points collected every 5 minutes during a midnight-to-midnight (24-hour) period - see Section III of Lobo et al. The daily CGM profiles from these 16 different studies formed three different datasets, each with a distinct purpose. 1. Training data composed of 23,916 daily CGM profiles sampled from the studies of DCLP1, DCLP3, DIAMOND1, DIAMOND2, DSS1, and NIGHTLIGHT, which was used to define a candidate set of CSCs. 2. Validation data composed of 37,758 daily CGM profiles sampled from the same 6 studies as the training data, which was used to evaluate the performance of the candidate set of CSCs and then select and fix the final set of CSCs. 3. Test data composed of 143,036 daily CGM profiles taken from the studies of CITY, DCLP5, DIAMOND2, NDIAB, MDEX, REPLACE - BG, RT - CGM, SENCE, SEVHYPO, UVA - TRIALNET, and WISDM, which was used to evaluate the robustness and generalizability of the final selected set of CSCs.

[0157] These datasets were constructed such that there was no overlap between the training data, validation data, and test data. Furthermore, the test data was derived from studies that used different methodologies and different generations of CGM technology compared to the studies used for training and validation data.

[0158] Analysis: Establishment of the final set of CSCs and classification of daily profiles Step 1: For any daily CGM profile, a set of CSCs is defined and then fixed by the property that there exists a CSC that approximates five standard time ranges of the daily CGM profile. The five standard time ranges are here abbreviated as follows: T54 (percent of CGM time below 54 mg / dl); T70 (percent of CGM time below 70 mg / dl), TIR (percent of CGM time within 70 - 180 mg / dl), T180 (percent of CGM time above 180 mg / dl), and T250 (percent of CGM time above 250 mg / dl). Such approximation ensures that the main clinically relevant properties of the daily CGM profile are preserved. The CSC set is defined using hierarchical clustering, where the input weighting is varied until the set of CSCs has the required performance to approximate the vector {T54, T70, TIR, T180, T250}.

[0159] Step 2: The procedure is developed to map the daily CGM profile to the CSC closest to it, including the step of computationally calculating a similarity measure (e.g., Euclidean distance) between the candidate daily CGM profile and the centroid of each CSC, and then the step of selecting the CSC with the best similarity measure value. The similarity measure is computationally calculated in the "space" defined by all possible vectors {T54, T70, TIR, T180, T250}. When these two steps are achieved, any daily CGM profile can be mapped to a CSC that approximates the five time ranges of the original daily CGM profile, and the sequence of an individual's CSCs can be used as a surrogate for the progression of this individual's blood glucose control.

[0160] Recalling that the glycemic control space is basically two-dimensional, by approximating the daily CGM profile by CSC, i.e., minimizing the distance between two points in {T54, T70, TIR, T180, T250}, any other measure of glycemic control derived from the daily CGM profile is similarly guaranteed to be approximated by CSC. In addition to measure approximation, the sequence of CSCs provides information regarding the timing and day-to-day variability of the patient's clinically relevant glycemic events. An extended mathematical explanation of the procedure described in this paragraph will be provided in the supplementary material.

[0161] Results Final Set of CSCs: Review and Interpretation The procedure described in the previous paragraph resulted in a final set of 35 clinically similar clusters. The results of this paragraph serve as an external validation of the CSC method using the training set and as an illustration of the potential for clinical application. The CSC index indicates the degree of glycemic control represented by TIR alone. TIR is highest in CSCs with low indices, such as 1, 2, 3, and lowest in CSCs with the highest indices, such as 33, 34, 35. For example, CSC1 has a TIR = 85.4%, and CSC35 has a TIR = 2.4%. However, CSCs with adjacent indices and similar TIRs can be very different in terms of exposure to hypoglycemia or hyperglycemia. For example, in CSC12, TIR = 46.4%, and the rest is distributed between levels 1 and 2 hypoglycemia, with a large presence of readings below 54 mg / dl (T54 = 27.3%). In contrast, in CSC13, TIR = 44.9%, but the rest is occupied by hyperglycemia (T180 = 30.4% and T250 = 21.9%).

[0162] Table 8 enumerates all CSCs with respective values of {T54, T70, TIR, T180, T250}, i.e., the centroid defines each CSC, and the number / percentage of the daily CGM profile is associated with each CSC of the test data. Table 8: List of All CSCs with Respective Values of {T54, T70, TIR, T180, T250} JPEG2025516193000177.jpg167158

[0163] Establishment of the final set of CSCs and classification of daily profiles The time ranges {T54, T70, TIR, T180, T250} were used as the input for a single daily CGM profile when performing hierarchical clustering. The input was generated using all daily CGM profiles in the training data. The implementation of hierarchical clustering of the scipy.cluster.hierarchy Python module using the centroid algorithm 17 was used to calculate the Euclidean distance between two rows of the input. Since the inventors wanted the hypoglycemic behavior to be faithfully captured by the CSCs, the inventors weighted the T54 and T170 input columns more heavily than the TIR, T180, and T250 input columns. Each CSC is a collection of daily CGM profiles such that each daily CGM profile in the collection has basically the same time range. For a given set of inputs (determined using the daily CGM profiles in the training data and the weights selected for the ultra-low and low columns), the hierarchical clustering algorithm generated a dendrogram indicating the hierarchical relationship between the daily CGM profiles in the training data.

[0164] "Cutting" the dendrogram at a specific height generated clustering by a specific number of clusters. The evaluation of each candidate set of CSCs used the validation data. The centroid of the CSCs was used to classify each daily CGM profile of the validation data. Let JPEG2025516193000178.jpg711 be the time range k of the i-th daily CGM profile from individual i, and let JPEG2025516193000179.jpg823 be the time range k of the CSC into which the j-th daily CGM profile from individual i is classified, where JPEG2025516193000180.jpg762.

[0165] Furthermore, Let JPEG2025516193000181.jpg be 1941, Let JPEG2025516193000182.jpg be 1866, where j is the set of all daily CGM profiles of an individual JPEG2025516193000183.jpg is the set of all 7150 daily CGM profiles. The single "optimal" set of CSCs.

[0166] Each k The set of points for maximizes the r-squared value of the linear regression through JPEG2025516193000184.jpg736, and the absolute value of the relative effect size is less than 0.15 for each k where the relative effect size is computer-calculated as the mean value of the mean difference between JPEG2025516193000185.jpg712 divided by the standard deviation of the whole JPEG2025516193000186.jpg712 and JPEG2025516193000187.jpg1126.

[0167] The inventors explored nine different weightings (i.e., nine different sets of inputs to the hierarchical clustering algorithm) of the ultra-low and low input columns resulting in a set of 166 candidate CSCs to be evaluated. The final fixed set of CSCs had 35 clusters. Each CSC is defined by its centroid. The centroid of a given CSC is calculated using the daily CGM profiles in the training data assigned to the CSC.

[0168] "Severe hypoglycemia" CSCs such as #10 and #22 are rarely seen (54 and 46 times respectively from the 143,036 daily CGM profiles), and "moderate" clusters such as #6 and #11 (6,770 and 1,294 times respectively), or clusters predominantly interspersed with hyperglycemia, e.g., #34 and #35 (4,226 and 1,146 times respectively), are frequent, which is clear when considering that the majority of the daily CGM profiles in this analysis come from individuals with variously treated T1D.

[0169] The most frequent CSCs are #1 and #5 by the large digits of a function that includes healthy people, and people on advanced treatments such as T1D on AID, or T2D on CGM (Table 8).

[0170] Tracking the progression of blood glucose control over time Over time, the sequence of daily CGM profiles generated by each individual can be presented as a sequence of CSCs that represents the progression of this individual's blood glucose control over time. For example, individuals with good blood glucose control visit fewer distinct CSCs and generally have a low index (indicating more time spent in range), while individuals with unstable blood glucose fluctuations visit more distinct CSCs, and the visited CSCs often have a high index (indicating less time spent in the target range and more time spent above or below the range). FIG. 14 is a three-panel plot showing the progression of three individuals with T1D over 14 days. Panel A (top row) presents data from a 6-year-old boy with a baseline HbA1c of 7.8% being treated by MDI from the SENCE study, Panel B (middle row) presents data from a 7-year-old boy with a baseline HbA1c of 7.9% during CGM+CSII treatment from the DCLP5 study, 27, and Panel C (bottom row) presents data from a 9-year-old boy with a baseline HbA1c of 7.8% during AID treatment from the DCLP5 study, representing these three treatment modalities. These individuals have the same gender, similar age, and essentially the same baseline HbA1c, but subsequently, these trajectories diverge and the daily transitions between the CSC and the number / index of visited CSCs are quite different.

[0171] Furthermore, Figure 14 includes the AGP for each of the presented 14-day CSC traces. It is clear that the CSC sequence faithfully represents the information carried by the AGP and adds information regarding the daily changes of an individual in glycemic control, these worst or best days, or any trends in the treatment progression that may occur during the 14-day observation. Even in the case of a person on AID treatment (Panel C) with very stable glycemic control having a TIR = 85.4% during the observation period, it becomes clear that there was one day with substantial hypoglycemia (day 8 classified as CSC#5), one hypoglycemic day (day 13, classified as CSC#2), and two days of incomplete data (days 2 and 12) when the AID system was broken or malfunctioned - information not conveyed by the AGP, or the TIR scale system.

[0172] CSC visits as an expression of health status and treatment effectiveness Figure 15 is a four-panel plot showing the ability of a set of CSCs to distinguish between health status and treatment modality. One panel represents the average number of unique CSCs (solid line) visited by people with T1D, people with T2D, and healthy individuals. It is clear that in T1D, the number of unique CSCs visited over time (e.g., 6 months) is the highest, while on average only three unique CSCs are visited by healthy individuals. Another panel shows the same trajectories as the first panel, but in this case distinguishing the treatments of individuals with T1D, namely MDI, CSII, and AID. The dotted lines in these panels are Weibull distribution functions that fit the data, and these fit curves approximate the actual trajectories accurately. The Weibull fit has a certain probabilistic meaning beyond the scope of this manuscript. Another panel shows box plots of the CSC index by health status for people with T1D separated by treatment modality. The plot confirms that on average the highest CSC index is reached by people with T1D during MDI treatment, with people with diabetes during AID treatment having the lowest CSC index, and healthy individuals visiting only a few CSCs with low indices (all below 10). Finally, another panel presents a statistical pentagram by Bonferroni-corrected pairwise comparisons between all five conditions under consideration. As shown in the figure, all pairwise differences are statistically significant except for the difference between T1D and T2D during CSII treatment.

[0173] Relationships between CSCs for blood glucose control, AGP, and established measures Each CSC represents the number of daily CGM profiles from different individuals. The relationship between the information carried by the CSC and the AGP is illustrated in FIG. 16A, which shows the two aforementioned adjacent clusters, #12 and #13. For each of these CSCs, instead of aggregating the individual's consecutive daily CGM profiles, the inventors plot a version of the AGP that aggregates all the daily CGM profiles associated with each CSC in this case. As expected, the AGP associated with CSC #13 is shifted up when compared to the AGP associated with CSC #12, while the AGP "clouds" are visually similar. This is expected to be quite different in terms of hypoglycemia and hyperglycemia even though CSCs #12 and #13 have similar TIRs (46.4% and 44.9% respectively). FIGS. 16B - 16J are plots similar to those shown in FIG. 16A for all 35 CSCs reported in this manuscript. Comparing the lowest to the highest CSC indices (e.g., #1, #2 vs. #34, #35) clearly shows the effect of good glycemic control on profiles mainly associated with hypoglycemia. Particularly beneficial are CSCs such as #28 or #32, which show a high variability in glycemic control with severe hypoglycemia and hyperglycemia.

[0174] To confirm the claim that the CSC faithfully represents the value of a commonly accepted measure of glycemic control in the CSC's member daily CGM profiles, Table 9 presents a univariate regression analysis with zero intercept. In each regression, the dependent variable is a measure computed from the daily CGM profile by computer, and the independent variable is the same measure computed from the CSC centroid associated with the daily CGM profile. As is clear from Table 9, the slope of all regressions approaches 1, indicating that the values of the measures computed from the CSC and the associated daily CGM profile are located close to the same line. Further, the R-squared values are high (generally above 0.81, corresponding to a correlation greater than 0.9), indicating that the measure computed from the CSC accounts for a good portion of the variability carried by the original daily CGM profile (with the exception of the coefficient of variation for reasons discussed). Table 9: Computer-Computed Measures of Glycemic Control from the Daily CGM Profile and Its Respective CSC TIFF2025516193000188.tif119153

[0175] Discussion After several years of fixation on a single measure of glycemic control, such as HbA1c, the medical community is increasingly recognizing that up-to-date technologies such as CGM provide a means of deep insight into a person's condition and the dynamics of glycemic variability. The system of TIR measures accepted by the international consensus of TIR17 and its extension to AID and clinical trial reports are excellent steps in this direction, provided that attention is paid to other quantifications such as the risk of hypoglycemia, as TIR alone does not become the only next marker of treatment effect. Simplifying further to a number of glycemic measures, it has been shown that two basic dimensions are sufficient to capture the information, hyperglycemic exposure, and risk of hypoglycemia carried by virtually all measures introduced to date.

[0176] One exception is the coefficient of variation, which has been shown to be incompatible with the TIR metric system and is generally controversial because of the fact that it is the ratio of two quantities (SD and mean) that are generally reduced by successful treatment for both. The TIR metric system adequately covers two dimensions by a certain mathematical redundancy, although adding up to 100% of the TIR component is clinically okay but may pose problems for some statistical methods. As a result, there is currently an established method for reviewing CGM data that includes the AGP and this adjacent TIR metric, and the inventors believe it to be robust enough to enable a framework for the classification and tracking of daily CGM profiles over time as they are further constructed.

[0177] In this manuscript, the inventors introduce a set of 35 clinically similar clusters that approximate the clinical impression conveyed by the original CGM data such that any daily CGM profile is assigned to a single CSC. Two attributes of the set of CSCs are important, namely (1) being finite and not overly large - at least in terms of the TIR metric, only 35 CSCs adequately classify the seemingly infinite number of daily CGM profiles at first glance, and (2) being fixed and not needing to be recalculated with new data. These attributes, finite and fixed, can form the basis for several clinical, computational, and algorithmic applications that essentially enable a table lookup solution to the clearly complex treatment optimization problem. A non-exhaustive list of potential applications includes the following. 1. Data structuring, dimensionality reduction, and database indexing: The continuum of all possible daily CGM profiles is clinically represented by the AGP and TIR scales and reduced to a finite and fixed set of CSCs that can be used as inputs to decision support, clinical, and automated treatment algorithms. A database indexed by the structure defined by the CSCs guarantees a fast and efficient search of subgroups of similar daily CGM profiles. This can facilitate features in decision support or AID systems such as algorithm learning from a person's CGM patterns and from other people's patterns stored in the database. 2. Distinction between health state and treatment modality: In another application of the CSCs, the inventors can imagine that a 10 - to 14 - day CGM wear in a home environment, likely with a predefined diet and physical activity schedule, can achieve diagnostic results similar to those acceptable in clinical practice based on the observed CSC patterns. Such an approach greatly simplifies data collection, replaces some common laboratory tests, and enables further remote treatment methods. 3. CGM pattern recognition and prediction: A transition probability matrix representing the occurrence of patients across the predefined set of CSCs is a natural [mathematically] tool for observing the progression of disease or treatment. Pattern recognition, or recurrent behavior, is reflected by patterns, or cycles, and detected by the transition probabilities from one CSC to the next. Predictions of short - term or long - term blood glucose control are based on recurrent visits to probabilistic patterns or a certain subset of CSCs. The latter is the subject of the theory of semi - Markov chains and results from the aggregation (ramping) of the state space to related subsets, characterized by the random durations consumed in each subset. 4. Tracking disease progression over time: Deterioration of glycemic control is indicated by a transition to an undesirable CSC, and conversely, successful treatment optimization or dosing titration is reflected by a transition to a clinically desirable CSC. In practical applications, the set of CSCs is labeled or ranked by clinical desirability, such as the TIR in Figure 1, and the label / rank of each CSC is fixed and used to steer treatment away from risk towards optimal control. Automation of the treatment process is facilitated by aggregating thousands of daily CGM profiles into a very small number of CSCs.

[0178] The main finding of this manuscript is that any daily CGM profile can be approximated by one of 35 pre - fixed clinical - similarity clusters. The approximation means that when a daily CGM profile is classified into a CSC, the CSC preserves the information carried by the original daily CGM profile in terms of the time - in - range scale system. Thus, the CSCs expand and, to some extent, complete the interpretation of CGM data provided by the AGP / TIR system. That is, if the AGP / TIR is a static snapshot of 14 - day data, the sequence of CSCs derived from the same data tracks the progression of glycemic control over time.

[0179] The time series of CSCs over 14 days indicates how stable or unstable a person's glycemic control is. In particular, this visualization allows us to examine days of an individual where glycemic control is quite different from what a person normally experiences. Since the set of CSCs condenses the clinical impressions conveyed by all possible daily CGM profiles into a finite, practical representation, it enables hosts of clinical applications, including database indexing, pattern recognition and tracking of treatment progression across a finite set of probabilities, decision support and table - lookup data interpretation for AID algorithms, or CGM exchange for common clinical tests.

[0180] FIG. 17 is an exemplary high-level functional block diagram of an embodiment of the present invention, or an aspect of an embodiment of the present invention. As shown in FIG. 17, a processor 104 or a controller communicates with a blood glucose monitor or data source 112, and optionally an insulin delivery device (e.g., another device 110). The blood glucose monitor or device communicates with a subject 1600 to monitor the blood glucose level of the subject 1600. The processor 104 or controller is configured to perform the required calculations. Optionally, the insulin delivery device communicates with the subject 1600 to deliver insulin to the subject 1600. The processor 104 or controller is configured to perform the required calculations. The blood glucose monitor and the insulin delivery device can be implemented as separate devices or as a single device. The processor 104 can be implemented locally in the blood glucose monitor, the insulin delivery device, or an independent device (or in any combination of two or more of the blood glucose monitor, the insulin device, or the independent device). A portion of the processor 104 or the system can be positioned remotely such that the device operates as a remote therapy device.

[0181] Referring to FIG. 18, in this most basic configuration, a computing device 1700 generally includes at least one processor 104 and a memory 106. Depending on the actual configuration and type of the computing device, the memory 106 can be volatile (such as RAM), non-volatile (such as ROM, flash memory), or a combination of both.

[0182] In addition, computing device 1700 can also have other features and / or functions. For example, computing device 1700 can also include additional removable and / or non-removable storage including, but not limited to, magnetic or optical disks or tapes, as well as writable electronic storage media. Such additional storage is in the form of removable storage 1702 and non-removable storage 1704. Computer storage media is implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data, including volatile and nonvolatile, removable and non-removable media. Memory, removable storage and non-removable storage are all examples of computer storage media. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CDROM, digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the required information and can be accessed by the device. Any such computer storage media can either be part of the device or used in conjunction with the device.

[0183] Computing device 1700 may also include one or more communication connections 1708 that enable the device to communicate with other devices (e.g., other computing devices). The communication connections carry information over a communication medium. The communication medium generally embodies computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery medium. The term "modulated data signal" means a signal that has one or more of its characteristics set or changed in such a manner as to encode, embody, or carry information in the signal. By way of example, and not limitation, communication media includes wired media such as a wired network or direct-wired connection, and wireless media such as radio, RF, infrared, and other wireless media. As noted above, the term computer readable media as used herein includes both storage media and communication media.

[0184] In addition to the independent computing machine, embodiments of the present invention can also be implemented in a network system including a plurality of computing devices that communicate with networking means such as a network by an infrastructure or an ad hoc network. The network connection can be a wired connection or a wireless connection. As an example, FIG. 18 shows a network system in which embodiments of the present invention can be implemented. In this example, the network system includes a computer 1706 (e.g., a network server), network connection means 1708 (e.g., wired and / or wireless connection), a computer terminal 1710, and a PDA (e.g., a smartphone) 1720 (or a mobile phone, a laptop computer, a tablet computer, a GPS receiver, an mp3 player, a handheld video player, a pocket projector, etc., or a handheld device (or a non-portable device) having a combination of such features). It should be understood that in one embodiment, module 1706 can be a blood glucose monitoring device. It should be understood that in one embodiment, the module listed as 1706 can be a blood glucose monitoring device, an artificial pancreas, and / or an insulin device (or other interventional or diagnostic device). Any of the components can be in a multiple number. Embodiments of the present invention can be implemented in any one of the devices of the system. For example, the execution of instructions or other required processing can be performed on the same computing device 1700. Alternatively, embodiments of the present invention can be executed on various computing devices of the network system. For example, a certain desired or required processing or execution can be performed on one of the computing devices of the network (e.g., server 1706 and / or blood glucose monitoring device), but other processing and execution of instructions can be performed on another computing device of the network system (e.g., terminal 1710), and vice versa.In practice, certain processing or execution can be performed on one computing device (e.g., server 1706 and / or an insulin device, an artificial pancreas, or a blood glucose monitoring device (or other interventional or diagnostic device)), and other processing or execution of instructions can be performed on different computing devices that may or may not be networked. For example, certain processing can be executed on terminal 1706, and other processing or instructions are passed to the computing device 1706 on which the instructions are to be executed. This scenario is particularly highly evaluated when a PDA device accesses the network, for example, via computer terminal 1710 (or an access point in an ad hoc network). In another embodiment, the protected software can be executed, encoded, or processed according to one or more embodiments of the present invention. The software that has then been processed, encoded, or executed can be distributed to customers. The distribution can be in the form of a storage medium (e.g., a disk) or an electronic copy.

[0185] FIG. 19 is a block diagram showing a system 100 including a connected computer system 1800 and an associated Internet 1802 that can implement an embodiment. Such a configuration is generally used for a computer (host) connected to the Internet 1802 and executing server or client (or a combination of) software. For example, source computers such as laptop computers, destination computers, and relay servers, as well as any computer or processor described herein, can use the computer system configuration and Internet connection shown in FIG. 19. System 1800 can be used as a portable electronic device such as a notebook / laptop computer, a media player (e.g., MP3-based or video player), a cellular phone, a personal digital assistant (PDA), a glucose monitoring device, an artificial pancreas, an insulin delivery device (or other interventional or diagnostic device), an image processing device (e.g., a digital camera or a video recorder), and / or any other handheld computing device, or any combination of these devices. FIG. 19 shows various components of a computer system, but it does not represent any particular architecture or manner of interconnecting the components. Note that such details are not closely related to the present invention. It will also be understood that network computers, handheld computers, cell phones, and other data processing systems with fewer components or perhaps more components can also be used. The computer system of FIG. 19 can be, for example, an Apple Macintosh computer or PowerBook, or an IBM-compatible PC. Computer system 100 includes a bus 1804, an interconnect, or other communication mechanism for transmitting information, and a processor 104 in the form of an integrated circuit generally coupled to the bus 1804 for processing information and executing computer-executable instructions. Computer system 100 also includes a main memory 106, such as a random access memory (RAM) or other dynamic storage device, coupled to the bus 1804 for storing information and instructions executed by the processor 104.

[0186] Main memory 106 can also be used to store temporary variables or other intermediate information during the execution of instructions executed by processor 104. Computer system 100 further includes a read-only memory (ROM) 136 (or other non-volatile memory) or other static storage device coupled to bus 1804 for storing static information and instructions of processor 104. Storage devices 1808, such as a magnetic disk or optical disk, a hard disk drive for reading and writing to a hard disk, a magnetic disk drive for reading and writing to a magnetic disk, and / or an optical disk drive (such as a DVD) for reading and writing to a removable optical disk, are coupled to bus 1804 for storing information and instructions. Hard disk drives, magnetic disk drives, and optical disk drives can be connected to the system bus by means of a hard disk drive interface, a magnetic disk drive interface, and an optical disk drive interface, respectively. The drives and the computer-readable media associated with the drives provide non-volatile storage of computer-readable instructions, data structures, program modules, and other data of a general-purpose computing device. Generally, computer system 100 includes an operating system (OS) stored in non-volatile storage for managing computer resources and providing access to computer resources and interfaces to applications and programs. The operating system generally processes system data and user input and responds by allocating and managing tasks and internal system resources, such as controlling and allocating memory, prioritizing system requests, controlling input and output devices, and facilitating networking and management of files. Non-limiting examples of operating systems are Microsoft Windows, Mac OS X, and Linux (registered trademark).

[0187] The term "processor" is meant to include, without limitation, any integrated circuit or other electronic device (or collection of devices) capable of executing the operations in at least one instruction, including reduced instruction set core (RISC) processors, CISC microprocessors, microcontroller units (MCU), CISC-based central processing units (CPU), and digital signal processors (DSP). The hardware of such devices can be integrated on a single substrate (e.g., a silicon "die") or distributed across two or more substrates. Furthermore, the various functional aspects of a processor can be implemented separately as software or firmware associated with the processor.

[0188] Computer system 100 can be coupled via bus 1804 to a display 1810, such as a cathode ray tube (CRT), liquid crystal display (LCD), flat screen monitor, touch screen monitor, or similar means for displaying text and graphical data to a user. The display can be connected via a video adapter for supporting the display. The display enables a user to view, input, and / or edit information related to the operation of the system. An input device 1812, including alphanumeric and other keys, is coupled to bus 104 for communicating information and command selections to processor 104. Another type of user input device is a cursor control 1814, such as a mouse, trackball, or cursor direction keys for communicating direction information and command selections to processor 104 and further controlling the movement of a cursor on display 1810. This input device generally has two degrees of freedom in two axes, a first axis (e.g., x) and a second axis (e.g., y), for indicating a position in a plane to the device.

[0189] Computer system 1800 can be used to implement the methods and techniques described herein. According to one embodiment, these methods and techniques are performed by computer system 1800 in response to processor 104 executing one or more sequences of one or more instructions embodied in main memory 1816. Such instructions can be read into main memory 106 from another computer-readable medium, such as storage device 1808. Execution of the sequence of instructions embodied in main memory 106 causes processor 104 to perform the processing steps described herein. In alternative embodiments, to implement the configuration, hardwired circuitry can be used in place of or in combination with software instructions. Accordingly, embodiments of the present invention are not limited to any specific combination of hardware circuitry and software.

[0190] As used herein, the term "computer-readable medium" (or "machine-readable medium") is an extensible term that refers to any medium or any memory that participates in providing instructions to a processor (such as processor 104) for execution, or any mechanism for storing or transmitting information in a form readable by a machine (e.g., a computer). Such a medium can store computer-executable instructions executed by a processing element and / or control logic, and data operated on by the processing element and / or control logic, and can further take many forms including, but not limited to, non-volatile media, volatile media, and transmission media. Transmission media include coaxial cables, copper wire, and fiber optics, including wires that include bus 1804. Transmission media can also take the form of acoustic or light waves, such as those generated during radio frequency and infrared data communications, or other forms of propagated signals (e.g., carrier waves, infrared signals, digital signals, etc.). Common forms of computer-readable media include, for example, floppy disks, flexible disks, hard disks, magnetic tape, or any other magnetic medium, CD-ROM, any other optical medium, punch cards, paper tape, any other physical medium having a pattern of holes, RAM, PROM, and EPROM, flash EPROM, any other memory chip or cartridge, carrier waves as described below, or any other medium readable by a computer.

[0191] Various forms of computer-readable media can be involved in transmitting one or more sequences of one or more instructions to a processor 104 that executes them. For example, the instructions can first be transmitted on a magnetic disk of a remote computer. The remote computer can load the instructions into its dynamic memory and transmit the instructions through a telephone line using a modem. A modem local to the computer system 100 can receive data on the telephone line and convert the data into an infrared signal using an infrared transmitter. An infrared detector can receive the data carried by the infrared signal, and an appropriate circuit can place the data on the bus 1804. The bus 1804 carries the data to the main memory 1816, and the processor 104 retrieves and executes the instructions from the main memory 1816. The instructions received by the main memory 1816 can optionally be stored in the storage device 1808 either before or after execution by the processor 104.

[0192] Computer system 100 also includes a communication interface 1818 coupled to bus 1804. Communication interface 1818 provides two-way data communication coupling to a network link 1822 connected to local network 1820. For example, communication interface 1818 can be an integrated services digital network (ISDN) card or modem and can provide a data communication connection to a corresponding type of telephone line. As another non-limiting example, communication interface 1818 can be a local area network (LAN) card for providing a data communication connection to a compatible LAN. For example, as described in "Ethernet Technology" on pages 7-1 through 7-38 of Chapter 7 of the "Internetworking Technology Handbook" with publication number 1-587005-001-3 (6 / 99) of Cisco Systems, incorporated herein by reference in its entirety for all purposes, Ethernet-based connections based on the IEEE 802.3 standard such as 10 / 100 Base T, 1000 Base T (Gigabit Ethernet), 10 Gigabit Ethernet (10GE or 10GbE or 10GigE according to the IEEE standard 802.3ae-2002 standard), 40 Gigabit Ethernet (40GbE), or 100 Gigabit Ethernet (100GbE according to the Ethernet standard IEEE 802.3ba) can be used. In such a case, communication interface 1818 generally includes a LAN transceiver or modem such as the Standard Microsystems Corporation (SMSC) LAN91C111 10 / 100 Ethernet transceiver described in the Standard Microsystems Corporation (SMSC) data sheet "LAN91C111 10 / 100 Non-PCI Ethernet Single-Chip MAC+PHY" data sheet, Revision 15 (02-20-04), incorporated herein by reference in its entirety for all purposes.

[0193] A wireless link can also be implemented. In any such implementation, communication interface 1818 transmits and receives electrical, electromagnetic, or optical signals that carry digital data streams representing various types of information.

[0194] Network link 1822 generally provides data communication to other data devices via one or more networks. For example, network link 1822 can provide a connection via local network 1822 to a host computer or to a data device operated by an Internet service provider (ISP) 1824. ISP 1824 provides data communication services via the worldwide packet data communication network Internet 1802. Both local network 1820 and Internet 1802 use electrical, electromagnetic, or optical signals that carry digital data streams. Signals via various networks and signals via communication interface 1818 that are on network link 1822 and carry digital data between computer system 100 are exemplary forms of carrier waves that transfer information.

[0195] The received code can be executed by the receiving processor 104 and / or stored in storage device 1808 or other non-volatile storage for later execution. Thus, computer system 100 can obtain application code in the form of a carrier wave.

[0196] The idea of identifying clinical similarity clusters of daily continuous glucose monitoring (CGM) profiles was developed by the inventor. The following steps of a) constructing and then fixing a set of clinical similarity clusters (CSCs) based on the property that there exists a clinical similarity cluster (CSC) that approximates the time-in-range of the daily CGM profile in any other daily continuous glucose monitoring (CGM) profile, and b) determining the approximation of any daily CGM profile by the CSCs developed by the inventor.

[0197] As can be seen from the requirements of the algorithms and methods discussed herein, the procedures are readily applicable to a device for identifying clinical similarity clusters of daily continuous glucose monitoring (CGM) profiles and can be implemented and utilized by associated processors, networks, computer systems, the Internet, and components and functions in accordance with the manner disclosed herein. As can be seen from the requirements of the algorithms and methods discussed herein, the procedures are readily applicable to a device for performing the steps of: a) constructing and then fixing a set of clinical similarity clusters (CSCs) based on the property that for any other daily continuous glucose monitoring (CGM) profile, there exists a clinical similarity cluster (CSC) that approximates the time-in-range of the daily CGM profile; and b) determining which daily CGM profiles are approximated by the CSCs, and can be implemented and utilized by associated processors, networks, computer systems, the Internet, and components and functions in accordance with the manner disclosed herein.

[0198] Figure 20 shows a system in which one or more embodiments of the present invention can be implemented using a network, or a portion of a network or computer. However, the blood glucose monitor, artificial pancreas or insulin device (or other interventional or diagnostic device) of the present invention can be implemented without a network. Figure 20 illustrates an exemplary system in which embodiments of the present invention can be implemented. In one embodiment, a glucose monitor, artificial pancreas or insulin device (or other interventional or diagnostic device) can be implemented locally at home or other desired location by a subject (or patient). However, in alternative embodiments, it can be implemented in a clinical setting or a support setting. For example, the clinic setup 1900 provides a place where a physician (e.g., 1902) or clinician / assistant diagnoses a patient (e.g., 1904) having a glucose-related disease and related disorders and conditions. The glucose monitoring device 1906 can be used as a stand-alone device to monitor and / or test a patient's glucose level. Although only the glucose monitor device 1906 is shown in the figure, it should be understood that the system of the present invention and any of its components can be used in the manner shown in Figure 20. The system or component can be attached to the patient or communicate with the patient if required or needed. For example, the system or a combination of components of the system - the glucose monitor device 1906 (or other associated devices or systems such as a controller, and / or an artificial pancreas, insulin pump (or other interventional or diagnostic device)), or any other desired or required device or component - can contact and communicate, or be fixed to the patient via a tape or tube (or other medical instrument or component), or communicate via a wired or wireless connection. Such monitoring and / or testing can be short-term (e.g., clinical visit) or long-term (e.g., clinical stay or family).The glucose monitoring device output can be used by a physician (clinician or assistant) for appropriate actions such as a patient's insulin injection or food intake, or other appropriate actions or modeling. Alternatively, the glucose monitoring device output can be sent to a computer terminal 1908 for immediate or future analysis. The transmission can be done via a cable, wirelessly, or any other appropriate medium. The glucose monitoring device output from the patient can also be sent to a portable device such as a PDA 1910. The glucose monitoring device output with improved accuracy can be sent to a glucose monitoring center 1912 for processing and / or analysis. Such transmission can be achieved in many ways such as a network connection 1914 which can be wired or wireless.

[0199] In addition to the glucose monitoring device output, errors, parameters for accuracy improvement, and any information regarding accuracy can be sent to a computer and / or a glucose monitoring center 1912 for error analysis. This can provide centralized accuracy monitoring, modeling, and / or accuracy expansion of a glucose center (or other interventional or diagnostic center) due to the importance of the glucose sensor (or other interventional or diagnostic sensor or device).

[0200] Embodiments of the present invention can also be implemented on an independent computing device associated with a target glucose monitoring device, an artificial pancreas, and / or an insulin device (or other interventional or diagnostic device).

[0201] FIG. 21 is a block diagram showing an example of a machine capable of implementing one or more aspects of an embodiment of the present invention. Referring to FIG. 21, aspects of an embodiment of the present invention include, but are not limited to, a system, method, and computer-readable medium that provide a step of identifying clinical similarity clusters of a daily continuous glucose monitoring (CGM) profile, and show a block diagram of an exemplary machine 2000 capable of implementing (e.g., executing) one or more embodiments (e.g., the methods discussed).

[0202] Referring to FIG. 21, aspects of an embodiment of the present invention include, but are not limited to, a system, method, and computer-readable medium that provide: a) a step of constructing and then fixing a set of clinical similarity clusters (CSCs) based on the property that for any other daily continuous glucose monitoring (CGM) profile, there exists a clinical similarity cluster (CSC) that approximates the time-in-range of the daily CGM profile; and b) a step of determining which daily CGM profile is approximated by a CSC, and show a block diagram of an exemplary machine 2000 capable of implementing (e.g., executing) one or more embodiments (e.g., the methods discussed).

[0203] An embodiment of machine 2000 can include logic, one or more components, circuits (e.g., modules), or mechanisms. A circuit is a tangible entity configured to perform certain operations. In one embodiment, circuits can be arranged in an instructed manner (e.g., inside an external entity such as another circuit or to an external entity such as another circuit). In one embodiment, one or more computer systems (e.g., stand-alone, client, or server computer systems) or one or more hardware processors (processors) can be configured by software (e.g., instructions, a portion of an application, or an application) as a circuit that operates to perform certain operations described herein. In one embodiment, the software can be present in (1) a non-transitory machine-readable medium or (2) a transmitted signal. In one embodiment, the software causes the circuit to perform certain operations when executed by the original hardware of the circuit.

[0204] In one embodiment, a circuit can be implemented mechanically or electronically. For example, a circuit can include dedicated circuits or logic specifically configured to perform one or more of the technologies described above, including a dedicated processor, a field programmable gate array (FPGA), or an application specific integrated circuit (ASIC). In one embodiment, a circuit can include programmable logic (e.g., a circuit included within a general purpose processor or other programmable processor) that can be temporarily configured (e.g., by software) to perform certain operations. It will be appreciated that the decision to implement a circuit mechanically (e.g., in dedicated and persistent circuits) or in a temporarily configured circuit (e.g., configured by software) is made based on cost and time considerations.

[0205] Accordingly, the term "circuit" is understood to include a tangible entity that can be physically constructed, permanently configured (e.g., hardwired), or temporarily (e.g., for a short time) configured (programmed) to operate in the indicated manner or to perform the indicated operations. In an example, given a plurality of temporarily configured circuits, each circuit need not be configured or initialized at any one instant in time. For example, if a circuit includes a general-purpose processor configured via software, the general-purpose processor can be configured as respective different circuits at different times. Thus, software can, for example, configure a processor to construct a particular circuit at one instant and different circuits at different instants.

[0206] In one example, a circuit can provide information to and receive information from other circuits. In this example, the circuit is considered to be communicatively coupled to one or more other circuits. If a plurality of such circuits are present simultaneously, communication can be achieved via signal transmission that connects the circuits (e.g., through appropriate circuits and buses). In embodiments where a plurality of circuits are configured or initialized at different times, such communication between circuits can be achieved, for example, via storage and retrieval of information in a memory structure to which the plurality of circuits have access. For example, one circuit can perform an operation and store the output of this operation in a communicatively coupled memory device. Then another circuit can access the memory device at a later time to retrieve and process the stored output. In one example, a circuit can be configured to initiate or receive communication with an input or output device and operate on resources (e.g., a collection of information).

[0207] The various operations of the examples of the methods described herein can be performed, at least in part, by one or more processors that are temporarily configured (e.g., by software) to perform the associated operations or are permanently configured. Whether temporarily or permanently configured, such processors can construct processor-implemented circuits that operate to perform one or more operations or functions. In one example, the circuits shown herein can include processor-implemented circuits.

[0208] Similarly, the methods described herein can be performed, at least in part, on a processor. For example, at least some of the operations of the method can be performed by one or more processors or processor-implemented circuits. Certain operations of the performance can be distributed among one or more processors and can exist within a single machine as well as be deployed across several machines. In one example, one or more processors can be located in a single location (e.g., in a home environment, in an office environment, or as a server farm), and in other examples, the processors can be distributed across several locations.

[0209] One or more processors can also operate to assist in the performance of the associated operations in a “cloud computing” environment or to operate as “software as a service” (SaaS). For example, at least some of the operations can be performed by a group of computers (as an example of machines including processors), and these operations can be accessible via a network (e.g., the Internet) and via one or more appropriate interfaces (e.g., an application program interface (API)).

[0210] Exemplary embodiments (e.g., apparatus, system, or method) can be implemented in digital electronic circuitry, in computer hardware, firmware, software, or in any combination thereof. Exemplary embodiments can be implemented using a computer program product (i.e., a computer program tangibly embodied in an information carrier or machine-readable medium to be executed by, or to control the operation of, a data processing apparatus such as a programmable processor, a computer, or multiple computers).

[0211] The computer program can be written in any form of programming language, including compiled or interpreted languages, and can be deployed in any form, including as a stand-alone program or as a software module, subroutine, or other unit suitable for use in a computing environment. The computer program can be deployed to be executed on one computer or on multiple computers at one site or can be distributed across multiple sites interconnected by a communication network.

[0212] In one embodiment, operations can be performed by one or more programmable processors executing a computer program that operates on input data to produce output. Embodiments of the method operations can also be performed by an exemplary apparatus, and the exemplary apparatus can be implemented as special purpose logic circuitry (e.g., a field programmable gate array (FPGA) or an application specific integrated circuit (ASIC)).

[0213] A computing system can include a client and a server. The client and server are generally remote from each other and typically interact via a communication network. The relationship between the client and server results from computer programs executed on respective computers and have a client-server relationship with each other. In embodiments deploying a programmable computing system, it will be appreciated that both hardware and software architectures need to be considered. In particular, it will be appreciated that a choice can be made in design selection as to whether to implement certain functions with permanently configured hardware (e.g., ASIC), with temporarily configured hardware (e.g., a combination of software and a programmable processor), or with a combination of permanently and temporarily configured hardware. The following shows hardware (e.g., machine 2000) and software architectures that can be deployed in exemplary embodiments.

[0214] In one example, machine 2000 can operate as a stand-alone device or machine 2000 can be connected (e.g., networked) to other machines.

[0215] In a networked deployment, machine 2000 can operate with the capacity of either a server or a client machine in a server-client network environment. In one example, machine 2000 can act as a peer machine in a peer-to-peer (or other distributed) network environment. Machine 2000 can be a personal computer (PC), tablet PC, set-top box (STB), personal digital assistant (PDA), mobile phone, web appliance, network router, switch or bridge, or any machine capable of executing instructions (sequential or otherwise) that direct the actions taken (e.g., executed) by machine 2000. Further, although only a single machine 2000 is shown, the term "machine" shall be construed to include any collection of machines that individually or jointly execute a set (or multiple sets) of instructions to perform any one or more of the methodologies discussed herein.

[0216] An exemplary machine (e.g., a computer system) 2000 can include a processor 104 (e.g., a central processing unit (CPU), a graphics processing unit (GPU), or both), main memory 106, and static memory 106, some or all of which can communicate with each other via bus 2020. Machine 2000 can further include a display unit 2002, alphanumeric input device 2004 (e.g., a keyboard), and a user interface (UI) navigation device 2006 (e.g., a mouse). In one example, display unit 2002, input device 2004, and UI navigation device 2006 can be a touch screen display. Machine 2000 can additionally include a storage device (e.g., a drive unit) 2008, signal generation device 2010 (e.g., a speaker), network interface device 2012, and one or more sensors 2014, such as a global positioning system (GPS) sensor, compass, accelerometer, or other sensor.

[0217] The storage device 2008 can include a machine-readable medium 2016 and store one or more sets of data structures or instructions 108 (e.g., software) that are implemented or utilized by any one or more of the methods or functions described herein. The instructions 108 can also be fully or at least partially present in the main memory 106, in the static memory 106, or in the processor 104 during execution by the machine 2000. In one embodiment, one or any combination of the processor 104, the main memory 106, the static memory 106, or the storage device 2008 can construct a machine-readable medium.

[0218] Although the machine-readable medium 2016 is shown as a single medium, the term "machine-readable medium" can include a single medium or a plurality of media (e.g., centralized or distributed databases, and / or associated caches and servers) configured to store one or more instructions 108. The term "machine-readable medium" can also be understood to include any tangible medium that can store, encode, or carry instructions for execution by a machine, cause the machine to perform any one or more of the methods of the present disclosure, or store, encode, or carry data structures utilized by or associated with such instructions. The term "machine-readable medium" can thus be understood to include, without limitation, solid-state memories, and optical and magnetic media. Specific examples of machine-readable media can include, by way of example, semiconductor memory devices (e.g., electronically programmable read-only memories (EPROMs), electronically erasable programmable read-only memories (EEPROMs), and flash memory devices), magnetic disks such as internal hard disks and removable disks, magneto-optical disks, and CD-ROM and DVD-ROM disks, including non-volatile memories.

[0219] Furthermore, instruction 108 can be transmitted or received through communication network 2018 that uses a transmission medium via a network interface device using any one of several transfer protocols (e.g., frame relay, IP, TCP, UDP, HTTP, etc.). Exemplary communication networks can include, among others, local area networks (LANs), wide area networks (WANs), packet data networks (e.g., the Internet), mobile phone networks (e.g., cellular networks), plain old telephone service (POTS) networks, and wireless data networks (e.g., the IEEE 802.11 standard family known as Wi-Fi (registered trademark), the IEEE 802.16 standard family known as WiMax (registered trademark)), peer-to-peer (P2P) networks. The term "transmission medium" should be understood to include any intangible medium that can store, encode, or carry instructions for machine execution and includes digital or analog communication signals or other intangible media for facilitating such software communication.

[0220] Although exemplary embodiments of the present disclosure are described in detail herein as some examples, it should be understood that other embodiments are also contemplated. Accordingly, the present disclosure is not limited in scope to the details of the construction and arrangement of components shown in the following description or illustrated in the drawings. The present disclosure can implement other embodiments and can be implemented or achieved in various ways.

[0221] It should be understood that any element, part, section, subsection, or component described with respect to any of the above specific embodiments can be incorporated into, integrated with, or adapted to be used with any other embodiment described herein, unless otherwise specified in detail or unless the embodiment's device would have to be expressed as non-functional. Similarly, any step described with respect to a particular method or process can be integrated into, incorporated into, or otherwise combined with other methods or processes described herein, unless otherwise specified in detail or unless the embodiment's method would have to be expressed as non-functional. Furthermore, the devices of multiple embodiments or the methods of embodiments can be combined with, incorporated into, or otherwise integrated with each other to construct or develop other embodiments of the invention described herein.

[0222] It should be understood that any of the components or modules shown with respect to any of the embodiments of the invention discussed herein can be formed integrally with or separately from each other. Furthermore, redundant functions or structures of the components or modules can be implemented. Additionally, the various components can communicate locally and / or remotely with any user / physician / patient or machine / system / computer / processor. Additionally, the various components can communicate via wireless and / or hardwire or other desired and available communication means, systems, and hardware. Additionally, the various components and modules can be replaced with other modules or components that provide similar functionality.

[0223] It should be understood that the devices and associated components discussed herein can accept all shapes along the entire continuous geometric spectrum of operations in the x, y, and z planes in order to provide and satisfy physical, environmental, and structural requirements and operating requirements. Additionally, the positions and arrangements of the various components can be changed if required or necessary.

[0224] It should be understood that the various sizes, dimensions, outer shapes, rigidities, shapes, flexibilities, and materials of any of the components or portions of the components in the various embodiments discussed throughout can be changed and utilized if required or necessary.

[0225] Although some dimensions are shown in the figures described above, it should be understood that the device can be configured with various sizes, dimensions, outer shapes, rigidities, shapes, flexibilities, and materials when related to a component or portion of the component of the device, and thus can be changed and utilized if required or necessary.

[0226] Also, it should be noted that the singular forms of "a", "an", and "the" used in the specification and the appended claims include plural references unless the context clearly dictates otherwise. A range can be expressed herein as "about" or "approximately" from one particular value and / or to "about" or "approximately" another particular value. When such a range is expressed, other exemplary embodiments include from one particular value and / or to another particular value.

[0227] By "comprising" or "including" or "containing" is meant that what is named, at least a composite, element, particle, or method step, is present in the composition or article or method, but does not exclude the presence of other such composites, materials, particles, or method steps having the same function as what is named.

[0228] When describing exemplary embodiments, we will borrow the help of terms for clarity. Each term is intended to have the broadest meaning understood by those skilled in the art and shall include all technical equivalents that operate in a similar manner to achieve similar purposes. Also, it should be understood that the mention of one or more steps of a method does not exclude the presence of additional method steps or intervening method steps between the explicitly identified steps. The steps of the method can be executed in an order different from that described herein without departing from the scope of the present disclosure. Similarly, it should be understood that the mention of one or more components in a device or system does not exclude the presence of additional components or intervening components between these explicitly identified components.

[0229] Some citations, which can include various patents, patent applications, and publications, are mentioned in a citation list and discussed in the disclosure provided herein. The mention and / or discussion of such citations are merely presented to clarify the description of the present disclosure and do not constitute an admission that any such citation is "prior art" with respect to any aspect of the present disclosure described herein. For annotation purposes, "[n]" corresponds to the nth citation in the list. All citations mentioned and discussed in this specification are incorporated by reference to the same extent as if each citation were incorporated by reference in its entirety and individually.

[0230] It should be understood that the subject matter discussed herein can be done to a human or any animal. The animal can be of any of a variety of applicable types, including but not limited to mammals, veterinary animals, livestock animals, or pet - type animals. By way of example, the animal can be a laboratory animal (e.g., rat, dog, pig, ape) specifically selected to have certain characteristics similar to humans. It should be understood that the subject can be done to any applicable human patient, for example.

[0231] As used herein, the term "about" means approximately, around, roughly, or generally. When the term "about" is used with a numerical range, the range is modified by extending the boundaries above and below the recited numerical values. In general, the term "about" is used herein to modify a numerical value by a variance of 10% above and below the stated value. In one aspect, the term "about" means plus or minus 10% of the numerical value of the number being used. Thus, about 50% means a range from 45% to 55%. It should also be understood that all numerical ranges recited herein by endpoints include all numbers and fractions subsumed within that range (e.g., 1 to 5 includes 1, 1.5, 2, 2.75, 3, 3.90, 4, 4.24, and 5). Similarly, all numerical ranges recited herein by endpoints include sub-ranges subsumed within that range (e.g., 1 to 5 includes 1-1.15, 1.5-2, 2-2.75, 2.75-3, 3-3.90, 3.90-4, 4-4.24, 4.24-5, 2-5, 3-5, 1-4, and 2-4). It should also be understood that all numbers and fractions are assumed to be modified by the term "about".

[0232] Additional description of aspects of the present disclosure is shown with reference to the accompanying drawings. The drawings form a part of a particular embodiment or example and illustrate a particular embodiment or example by way of illustration.

[0233] The following references are hereby incorporated by reference in their entirety. 1.M. D. Breton and B. P. Kovatchev, "One year real-world use of the control-IQ advanced hybrid closed-looPtechnology," Diabetes technology & therapeutics, vol. 23, p. 601-608, 2021. 2. B. P. Kovatchev, "Metrics for glycaemic control - from HbA 1c to continuous glucose monitoring," Nature Reviews Endocrinology, vol. 13, p. 425 - 436, 2017. 3. B. Kovatchev and C. Cobelli, "Glucose variability: timing, risk analysis, and relationshiPto hypoglycemia in diabetes," Diabetes Care, vol. 39, p. 502 - 510, 2016. 4. C. Cobelli, C. Dalla Man, G. Sparacino, L. Magni, G. De Nicolao and B. P. Kovatchev, "Diabetes: models, signals, and control," IEEE reviews in biomedical engineering, vol. 2, p. 54 - 96, 2009. 5. B. P. Kovatchev, W. L. Clarke, M. Breton, K. Brayman and A. McCall, "Quantifying temporal glucose variability in diabetes via continuous glucose monitoring: mathematical methods and clinical application," Diabetes technology & therapeutics, vol. 7, p. 849 - 862, 2005. 6.C. M. McDonnell, S. M. Donath, S. I. Vidmar, G. A. Werther and F. J. Cameron, "A novel approach to continuous glucose analysis utilizing glycemic variation," Diabetes technology & therapeutics, vol. 7, p. 253-263, 2005. 7.P. A. Baghurst, "Calculating the mean amplitude of glycemic excursion from continuous glucose monitoring data: an automated algorithm," Diabetes technology & therapeutics, vol. 13, p. 296-302, 2011. 8.C. Fabris, S. D. Patek and M. D. Breton, "Are risk indices derived from CGM interchangeable with SMBG-based indices?," Journal of diabetes science and technology, vol. 10, p. 50-59, 2016. 9.A. L. McCall, D. J. Cox, J. Crean, M. Gloster and B. P. Kovatchev, "A novel analytical method for assessing glucose variability: using CGMS in type 1 diabetes mellitus," Diabetes technology & therapeutics, vol. 8, p. 644-653, 2006. 10.D. Rodbard, "New and improved methods to characterize glycemic variability using continuous glucose monitoring," Diabetes technology & therapeutics, vol. 11, p. 551-565, 2009. 11.D. Rodbard, "Interpretation of continuous glucose monitoring data: glycemic variability and quality of glycemic control," Diabetes technology & therapeutics, vol. 11, p. S-55, 2009. 12.W. Clarke and B. Kovatchev, "Statistical tools to analyze continuous glucose monitor data," Diabetes technology & therapeutics, vol. 11, p. S-45, 2009. 13.L. Magni, D. M. Raimondo, C. D. Man, M. Breton, S. Patek, G. De Nicolao, C. Cobelli and B. P. Kovatchev, "Evaluating the efficacy of closed-looPglucose regulation via control-variability grid analysis," Journal of diabetes science and technology, vol. 2, p. 630-635, 2008. 14. T. Battelino, T. Danne, R. M. Bergenstal, S. A. Amiel, R. Beck, T. Biester, E. Bosi, B. A. Buckingham, W. T. Cefalu, K. L. Close and others, "Clinical targets for continuous glucose monitoring data interpretation: recommendations from the international consensus on time in range," Diabetes care, vol. 42, p. 1593 - 1603, 2019. 15. R. S. Mazze, D. Lucido, O. Langer, K. Hartmann and D. Rodbard, "Ambulatory glucose profile: representation of verified self - monitored blood glucose data," Diabetes Care, vol. 10, p. 111 - 117, 1987. 16. R. M. Bergenstal, A. J. Ahmann, T. Bailey, R. W. Beck, J. Bissen, B. Buckingham, L. Deeb, R. H. Dolin, S. K. Garg, R. Goland and others, Recommendations for standardizing glucose reporting and analysis to optimize clinical decision making in diabetes: the Ambulatory Glucose Profile (AGP), Mary Ann Liebert, Inc. 140 Huguenot Street, 3rd Floor New Rochelle, NY 10801 USA, 2013. 17. A. D. Association, "7. Diabetes Technology: Standards of Medical Care in Diabetes - 2019," Diabetes Care, vol. 42, pp. S71 - S80, December 2018. 18. V. Fonseca and G. Grunberger, "Letter to the Editor," Endocrine Practice, vol. 23, pp. 629 - 632, 2017. 19. B. Kovatchev, S. Anderson, D. Raghinaru, Y. Kudva, L. a. L. C. Laffel, J. Pinsker and others, "Randomized Controlled Trial of Mobile Closed - Loop Control," Diabetes Care, vol. 43, no. 3, pp. 607 - 615, 2020. 20. S. Brown, B. Kovatchev, D. Raghinaru, J. Lum, B. Buckingham, Y. Kudva, L. Laffel and others, "Six - Month Randomized, Multicenter Trial of Closed - Loop Control in Type 1 Diabetes," New England Journal of Medicine, vol. 381, no. 18, pp. 1707 - 1717, 2019. 21.R. Beck, T. Riddlesworth, K. Ruedy, A. Ahmann, R. Bergenstal, S. Haller, C. Kollman and others, "Effect of Continuous Glucose Monitoring on Glycemic Control in Adults With Type 1 Diabetes Using Insulin Injections: The DIAMOND Randomized Clinical Trial," Journal of the American Medical Association, vol. 317, no. 4, pp. 371-378, 2017. 22.R. Beck, T. Riddlesworth, K. Ruedy, A. Ahmann, S. Haller, D. Kruger, J. McGill and others, "Continuous Glucose Monitoring Versus Usual Care in Patients With Type 2 Diabetes Receiving Multiple Daily Insulin Injections," Annals of Internal Medicine, vol. 167, no. 6, pp. 365-374, 2017. 23.A. Bisio, S. Anderson, L. Norlander, G. O'Malley, J. Robic, S. Ogyaadu, L. Hsu and others, "Impact of a Novel Diabetes Support System on a Cohort of Individuals With Type 1 Diabetes Treated With Multiple Daily Injections: A Multicenter Randomized Study," DIabetes Care, vol. 45, no. 1, pp. 186-193, 2022. 24.B. Kovatchev, L. Kollar, S. Anderson, C. Barnett, M. Breton, K. Carr, R. Gildersleeve and others, "Evening and overnight closed-looPcontrol versus 24 / 7 continuous closed-looPcontrol for type 1 diabetes: a randomised crossover trial," The Lancet Digital Health, vol. 2, no. 2, pp. e64-e73, 2020. 25.L. Laffel, L. Kanapka, R. Beck, K. Bergamo, M. Clements, A. Criego, D. DeSalvo and others, "Effect of Continuous Glucose Monitoring on Glycemic Control in Adolescents and Young Adults With Type 1 Diabetes: A Randomized Clinical Trial," Journal of the American Medical Association, vol. 323, no. 23, pp. 2388-2396, 2020. 26.M. Breton, L. Kanapka, R. Beck, L. Ekhlaspour, G. Forlenza, E. Cengiz, M. Schoelwer and others, "A randomized trial of closed-looPcontrol in children with type 1 diabetes," New England Journal of Medicine, vol. 383, no. 9, pp. 836-845, 2020. 27.V. Shah, S. DuBose, Z. Li, R. Beck, A. Peters, R. Weinstock, D. Kruger and others, "Continuous Glucose Monitoring Profiles in Healthy Nondiabetic Participants: A Multicenter Prospective Study," The Journal of Clinical Endocrinology & Metabolism, vol. 104, no. 10, pp. 4356-4364, 2019. 28.M. Rickels, S. DuBose, E. Toschi, R. Beck, A. Verdejo, H. Wolpert, M. Cummins and others, "Mini-Dose Glucagon as a Novel Approach to Prevent Exercise-Induced Hypoglycemia in Type 1 Diabetes," Diabetes Care, vol. 41, no. 9, pp. 1909-1916, 2018. 29.G. Aleppo, K. Ruedy, T. Riddlesworth, D. Kruger, A. Peters, I. Hirsch, R. Bergenstal and others, "REPLACE-BG: A Randomized Trial Comparing Continuous Glucose Monitoring With and Without Routine Blood Glucose Monitoring in Adults With Well-Controlled Type 1 Diabetes," Diabetes Care, vol. 40, no. 4, pp. 538-545, 2017. 30. JDRF CGM Study Group, "Continuous glucose monitoring and intensive treatment of type 1 diabetes," New England Journal of Medicine, vol. 359, no. 14, pp. 1464 - 1476, 2008. 31. L. Laffel, K. Harrington, A. Hanono, N. Naik, L. Ambler - Osborn, A. Schultz, L. DiMeglio and others, "A Randomized Clinical Trial Assessing Continuous Glucose Monitoring (CGM) Use With Standardized Education With or Without a Family Behavioral Intervention Compared With Fingerstick Blood Glucose Monitoring in Very Young Children With Type 1 Diabetes," Diabetes Care, vol. 44, no. 2, pp. 464 - 472, 2021. 32. R. Weinstock, S. DuBose, R. Bergenstal, N. Chaytor, C. Peterson, B. Olson, M. Munshi and others, "Risk Factors Associated With Severe Hypoglycemia in Older Adults With Type 1 Diabetes," Diabetes Care, vol. 39, no. 4, pp. 603 - 610, 2016. 33. A. Carlson, L. Kanapka, K. Miller, A. Ahmann, N. Chaytor, S. Fox, L. Kiblinger and others, "Hypoglycemia and Glycemic Control in Older Adults With Type 1 Diabetes: Baseline Results From the WISDM Study," Journal of Diabetes Science and Technology, vol. 15, no. 3, pp. 582 - 592, 2021. 34. B. Lobo, L. Farhy, M. Shafiei and B. Kovatchev, "A data-driven approach to classifying daily continuous glucose monitoring (CGM) time series," IEEE Transactions on Biomedical Engineering, vol. 62, no. 2, pp. 654 - 665, 2022. 35. P. Virtanen, R. Gommers, T. Oliphant, M. Haberland, T. Reddy, D. Cournapeau and others, "SciPy 1.0: Fundamental Algorithms for Scientific Computing in Python," Nature Methods, vol. 17, pp. 261 - 272, 2020. 36. National Institute of Standards and Technology, "Weibull Distribution," [Online]. Available: https: / / www.itl.nist.gov / div898 / handbook / eda / section3 / eda3668.htm. [Accessed 13 12 2021]. 37. International Patent Application Serial No. PCT / US2020 / 032121, titled "Methods for Constructing and Classifying Continuous Glucose Monitoring (CGM) Profiles", filed on June 3, 2022. 38. U.S. Utility Patent Application Serial No. 17 / 829,754, titled "Methods for Constructing and Classifying Continuous Glucose Monitoring (CGM) Profiles", filed on June 1, 2022. 39. U.S. Utility Patent Application Serial No. 17 / 776,717, titled "Model-Based Tracking of Hemoglobin A1c from Daily Continuous Glucose Monitoring Profiles and Systems", filed on May 13, 2022. 40. International Patent Application Serial No. PCT / US2020 / 060658, titled "Model-Based Tracking of Hemoglobin A1c from Daily Continuous Glucose Monitoring Profiles and Systems", filed on November 16, 2020; Publication No. WO2021 / 097396, on May 20, 2021. 41. U.S. Utility Patent Application Serial No. 17 / 776,056, titled "Systems, Methods, and Computer-Readable Media for Compressing Continuous Glucose Monitor Data", filed on May 11, 2022. 42. International Patent Application Serial No. PCT / US2020 / 060366, titled "Systems, Methods, and Computer-Readable Media for Compressing Continuous Glucose Monitor Data", filed on November 13, 2020; Publication No. WO / 2021 / 097178, on May 20, 2021. 43. U.S. Utility Patent Application Serial No. 17 / 694,870, titled "Insulin Monitoring and Delivery Systems and Methods for CGM-Based Disorder Detection and Mitigation via Tracking of Metabolic State", filed on March 15, 2022; Publication No. US2022-0203020A1, on June 30, 2022. 44. U.S. Utility Patent Application Serial No. 15 / 580,935, titled "Insulin Monitoring and Delivery Systems and Methods for CGM-Based Disorder Detection and Mitigation via Tracking of Metabolic State", filed on December 8, 2017; U.S. Patent No. 11,311,665, issued on April 26, 2022. 45. International Patent Application Serial No. PCT / US2016 / 036729, titled "Insulin Monitoring and Delivery System and Method for CGM-Based Disorder Detection and Mitigation via Tracking of Metabolic State", filed on June 9, 2016; Publication No. WO2016 / 201120, on December 15, 2016. 46. U.S. Utility Patent Application Serial No. 17 / 683,676, titled "Systems, Methods, and Computer-Readable Media for Dynamically Tracking the Risk of Hypoglycemia in Type 1 and Type 2 Diabetes", filed on March 1, 2022; Publication No. US2022-0262519A1, on August 18, 2022. 47. U.S. Utility Patent Application Serial No. 15 / 958,257, titled "Systems, Methods, and Computer-Readable Media for Dynamically Tracking the Risk of Hypoglycemia in Type 1 and Type 2 Diabetes", filed on August 20, 2018; U.S. Patent No. 11,289,201, issued on March 29, 2022.

[0234] 48. International Patent Application Serial No. PCT / US2016 / 058234, titled "Systems, Methods, and Computer-Readable Media for Dynamically Tracking the Risk of Hypoglycemia in Type 1 and Type 2 Diabetes", filed on October 21, 2016; Publication No. WO2017 / 070553, on April 27, 2017. 49. International Patent Application Serial No. PCT / US2022 / 017489, titled "Methods and Systems for Quantitative Physiological Assessment and Prediction of Clinical Subtypes of Glucose Metabolism Disorders", filed on February 23, 2022; Publication No. WO / 2022 / 182736, on September 1, 2022. 50. International Patent Application Serial No. PCT / US2022 / 017449, titled "Methods and Systems for Mapping Individualized Metabolic Phenotypes to Database Images to Optimize Control of Chronic Metabolic States", filed on February 23, 2022; Publication No. WO2022182709, on September 1, 2022. 51. U.S. Utility Patent Application Serial No. 17 / 590,659, titled "Systems, Methods, and Computer Simulation Environments for In Silico Trials in Prediabetes and Type 2 Diabetes", filed on February 1, 2022; Publication No. US2022-0230762A1, on July 21, 2022. 52. U.S. Utility Patent Application Serial No. 13 / 380,839, titled "Systems, Methods, and Computer Simulation Environments for In Silico Trials in Prediabetes and Type 2 Diabetes", filed on December 25, 2011; U.S. Patent No. 11,238,990, issued on February 1, 2022. 53. International Patent Application Serial No. PCT / US2010 / 040097, titled "Systems, Methods, and Computer Simulation Environments for In Silico Trials in Prediabetes and Type 2 Diabetes", filed on June 25, 2010; Publication No. WO2010 / 151834, on December 29, 2010. 54. International Patent Application Serial No. PCT / US2021 / 045936, titled "Methods and Systems for Generating User-Adjustable Presentations of Glucose Homeostasis in Type 1 Diabetes Based on Automated Receipt of Treatment Profile Data", filed on August 13, 2021; Publication No. WO2022 / 036214, on February 17, 2022. 55. U.S. Utility Patent Application Serial No. 17 / 339,153, titled "Methods and Systems for Safety, Analysis, and Monitoring of Insulin Pump Action and Other Modes of Insulin Delivery in Diabetes", filed on June 4, 2021: Publication No. US2021-0313036A1, on October 7, 2021. 56. U.S. Utility Patent Application Serial No. 13,634,040, titled "Methods and Systems for Safety, Analysis, and Monitoring of Insulin Pump Action and Other Modes of Insulin Delivery in Diabetes", filed on September 11, 2012; U.S. Patent No. 11,069,434, issued on July 20, 2021. 57. International Patent Application Serial No. PCT / US2011 / 028163, titled "Methods and Systems for Safety, Analysis, and Monitoring of Insulin Pump Action and Other Modes of Insulin Delivery in Diabetes", filed on March 11, 2011; Publication No. WO2011 / 112974, on September 15, 2011. 58. U.S. Utility Patent Application Serial No. 17 / 333,161, titled "Methods, Systems, and Computer Program Products for CGM-Based Hypoglycemia Prevention via Assessment of Risk of Hypoglycemia and Reduction of Smooth Insulin Delivery", filed on May 28, 2021; Publication No. US2021-0282677A1, on September 16, 2021. 59. U.S. Utility Patent Application Serial No. 17 / 070,245, titled "Methods, Systems, and Computer Program Products for CGM-Based Hypoglycemia Prevention via Assessment of Risk of Hypoglycemia and Smooth Reduction of Insulin Delivery", filed on October 14, 2020; Publication No. US2021 / 0038132A1, on February 11, 2021. 60. U.S. Utility Patent Application Serial No. 15 / 669,111, titled "Methods, Systems, and Computer Program Products for CGM-Based Hypoglycemia Prevention via Assessment of Risk of Hypoglycemia and Smooth Reduction of Insulin Delivery", filed on August 4, 2017; U.S. Patent No. 10,842,419, issued on November 24, 2020. 61. U.S. Utility Patent Application Serial No. 14 / 015,831, titled "CGM-Based Hypoglycemia Prevention via Assessment of Risk of Hypoglycemia and Smooth Reduction of Insulin Delivery", filed on August 30, 2013; U.S. Patent No. 9,750,438, issued on September 5, 2017. 62. U.S. Utility Patent Application Serial No. 13 / 203,469, titled "CGM-Based Hypoglycemia Prevention via Assessment of Risk of Hypoglycemia and Smooth Reduction of Insulin Delivery", filed on August 25, 2011; U.S. Patent No. 8,562,587, issued on October 22, 2013. 63. International Patent Application Serial No. PCT / US2010 / 025405, titled "Evaluation of the Risk of Hypoglycemia and CGM-Based Hypoglycemia Prevention via Smooth Reduction of Insulin Delivery," filed on February 25, 2010; Publication No. WO2010 / 099313A1, on September 2, 2010. 64. U.S. Utility Patent Application Serial No. 17 / 181,888, titled "System for Centralized Data Exchange for Blood Glucose Monitoring and Control," filed on February 22, 2021; Publication No. US2021-0169409A1, on June 10, 2021. 66. U.S. Utility Patent Application Serial No. 15 / 109,682, titled "System for Centralized Data Exchange for Blood Glucose Monitoring and Control," filed on July 5, 2016; U.S. Patent No. 10,925,536, issued on February 23, 2021. 67. International Patent Application Serial No. PCT / US2015 / 010167, titled "Central Data Exchange Node for System Monitoring and Control of Blood Glucose Levels in Diabetic Patients," filed on January 5, 2015; Publication No. WO2015103543, on July 9, 2015. 68. U.S. Utility Patent Application Serial No. 17 / 180,301, titled "Impact on End-Stage Renal Disease Outcomes by Predicting Physiological Parameters and Determining Dosage Recommendations," filed on February 19, 2021; Publication No. US2021-0257096A1, on August 19, 2021. 69. U.S. Utility Patent Application Serial No. 17 / 156,169, titled "Method, System, and Computer Program Product for Detecting in Real Time a Drop in Sensitivity of an Analyte Sensor," filed on January 22, 2021; Publication No. US2021-0218481A1, on July 15, 2021. 70. U.S. Utility Patent Application Serial No. 15 / 866,384, titled "Method, System, and Computer Program Product for Detecting in Real Time a Drop in Sensitivity of an Analyte Sensor," filed on January 9, 2018; U.S. Patent No. 10,903,914, issued on January 26, 2021. 71. U.S. Utility Patent Application Serial No. 14 / 266,612, titled "Method, System, and Computer Program Product for Detecting in Real Time a Drop in Sensitivity of an Analyte Sensor", filed on April 30, 2014; U.S. Patent No. 9,882,660, issued on January 30, 2018. 72. U.S. Utility Patent Application Serial No. 13 / 418,305, titled "Method, System, and Computer Program Product for Detecting in Real Time a Drop in Sensitivity of an Analyte Sensor", filed on March 12, 2012; U.S. Patent No. 8,718,958, issued on May 6, 2014. 73. International Patent Application Serial No. PCT / US2007 / 082744, titled "Method, System, and Computer Program Product for Detecting in Real Time a Drop in Sensitivity of an Analyte Sensor", filed on October 26, 2007; Publication No. WO / 2008 / 052199, on May 2, 2008. 74. U.S. Utility Patent Application Serial No. 11 / 925,689, titled "Method, System, and Computer Program Product for Detecting in Real Time a Drop in Sensitivity of an Analyte Sensor", filed on October 26, 2007; U.S. Patent No. 8,135,548, issued on March 13, 2012. 75. U.S. Utility Patent Application Serial No. 17 / 132,604, titled "Method, System, and Device for Precision Continuous Glucose Monitoring", filed on December 23, 2020; Publication No. US2021-0113122A1, on April 22, 2021. 76. U.S. Utility Patent Application Serial No. 15 / 510,878, titled "Method, System, and Device for Precision Continuous Glucose Monitoring", filed on March 13, 2017; U.S. Patent No. 10,881,334, issued on January 5, 2021. 77. International Patent Application Serial No. PCT / US2015 / 045340, titled "Improved Method, System, and Device for Precision Continuous Glucose Monitoring", filed on August 14, 2015; Publication No. WO2016025874, on February 18, 2016. 78. U.S. Utility Patent Application Serial No. 16 / 789,901, titled "Unified Platform for Monitoring and Controlling Blood Glucose Levels in Diabetic Patients", filed on February 13, 2020; Publication No. US-2020-0214629-A1, on July 9, 2020. 79. U.S. Utility Patent Application Serial No. 14 / 128,922, titled "Unified Platform for Monitoring and Controlling Blood Glucose Levels in Diabetic Patients", filed on December 23, 2013; U.S. Patent No. 10,610,154, issued on April 7, 2020. 80. International Patent Application Serial No. PCT / US2012 / 043910, titled "Unified Platform for Monitoring and Controlling Blood Glucose Levels in Diabetic Patients", filed on June 23, 2012; Publication No. WO2012 / 178134, on December 27, 2012. 81. U.S. Utility Patent Application Serial No. 16 / 588,881, titled "Tracking the Probability of Imminent Hypoglycemia in Diabetes from Self-Monitoring Blood Glucose (SMBG) Data", filed on September 30, 2012; Publication No. US-2020-0066410-A1, on February 27, 2020. 82. U.S. Utility Patent Application Serial No. 13 / 394,091, titled "Tracking the Probability of Imminent Hypoglycemia in Diabetes from Self-Monitoring Blood Glucose (SMBG) Data", filed on March 2, 2012; U.S. Patent No. 10,431,342, issued on October 1, 2019. 83. International Patent Application Serial No. PCT / US2010 / 047711, titled "Tracking the Probability of Imminent Hypoglycemia in Diabetes from Self-Monitoring Blood Glucose (SMBG) Data", filed on September 2, 2010; Publication No. WO2011 / 028925, on March 10, 2011. 84. U.S. Utility Patent Application Serial No. 16 / 546,335, titled "System Coordinator and Modular Architecture for Open-Loop and Closed-Loop Control of Diabetes", filed on August 21, 2019; Publication No. US-2019-0374137-A1, on December 12, 2019. 85. U.S. Utility Patent Application Serial No. 13 / 322,943, titled "System Coordinator and Modular Architecture for Open Loop and Closed Loop Control of Diabetes", filed on November 29, 2011; U.S. Patent No. 10,420,489, issued on September 24, 2019. 86. International Patent Application Serial No. PCT / US2010 / 036629, titled "System Coordinator and Modular Architecture for Open Loop and Closed Loop Control of Diabetes", filed on May 28, 2010; Publication No. WO2010 / 138848, on December 2, 2010. 87. U.S. Utility Patent Application Serial No. 16 / 451,766, titled "Tracking Changes in Average Blood Glucose in Diabetes", filed on June 25, 2019; Publication No. US-2019-0318801-A1, on October 17, 2019. 88. U.S. Utility Patent Application Serial No. 14 / 769,638, titled "Methods and Systems for Model-Based Tracking of Changes in Average Blood Glucose in Diabetes", filed on August 21, 2015; U.S. Patent No. 10,332,615, issued on June 25, 2019.

[0235] 89. International Patent Application Serial No. PCT / US2014 / 017754, titled "Methods and Systems for Model-Based Tracking of Changes in Average Blood Glucose in Diabetes", filed on February 21, 2014; Publication No. WO / 2014 / 130841, on August 28, 2014. 90. U.S. Utility Patent Application Serial No. 17 / 180,301, titled "Impact on the Outcome of End-Stage Renal Disease by Predicting Physiological Parameters and Determining Administration Recommendations", filed on February 19, 2021; Publication No. US2021‐0257096A1, on August 19, 2021. 91. U.S. Utility Patent Application Serial No. 16 / 126,879, titled "Methods, Systems, and Computer Program Products for Evaluating Insulin Sensitivity, Insulin / Glucagon Ratio, and Insulin Correction Factor in Diabetes from Self-Monitoring Data", filed on September 10, 2018; Publication No. US-2019-0019571-A1, on January 17, 2019. 92. U.S. Utility Patent Application Serial No. 12 / 665,149, titled "Method, System, and Computer Program Product for Evaluating Insulin Sensitivity, Insulin / Carbohydrate Ratio, and Insulin Correction Factor in Diabetes from Self-Monitoring Data", filed on December 17, 2009; Publication No. 2010 / 0198520, on August 5, 2010. 93. International Patent Application Serial No. PCT / US2008 / 069416, titled "Method, System, and Computer Program Product for Evaluating Insulin Sensitivity, Insulin / Carbohydrate Ratio, and Insulin Correction Factor in Diabetes from Self-Monitoring Data", filed on July 8, 2008; Publication No. WO2009 / 009528, on January 15, 2009. 94. U.S. Utility Patent Application Serial No. 15 / 580,915, titled "System and Method for Tracking Changes in Average Blood Glucose in Diabetes", filed on December 8, 2017; Publication No. US-2018-0313815-A1, on November 1, 2018. 95. International Patent Application Serial No. PCT / US2016 / 036481, titled "System and Method for Tracking Changes in Average Blood Glucose in Diabetes", filed on June 8, 2016; Publication No. WO2016200970, on December 15, 2016. 96. U.S. Utility Patent Application Serial No. 14 / 902,731, titled "Simulation of Endogenous and Exogenous Glucose / Insulin / Glucagon Interactions in Type 1 Diabetic Patients", filed on January 4, 2016; U.S. Patent No. 10,169,544, issued on January 1, 2019. 97. International Patent Application Serial No. PCT / US2014 / 045393, titled "Simulation of Endogenous and Exogenous Glucose / Insulin / Glucagon Interactions in Type 1 Diabetic Patients", filed on July 3, 2014; Publication No. WO2015003124, on January 8, 2015. 98. U.S. Utility Patent Application Serial No. 14 / 799,329, titled "Accuracy of Continuous Glucose Sensors", filed on July 14, 2015; U.S. Patent No. 10,194,850, issued on February 5, 2019. 99. U.S. Utility Patent Application Serial No. 12 / 065,257, titled "Accuracy of Continuous Glucose Sensors", filed on February 28, 2008; Publication No. 2008 / 0314395, on December 25, 2008. 100. International Patent Application Serial No. PCT / US2006 / 033724, titled "Methods for Improving the Accuracy of Continuous Glucose Sensors and Continuous Glucose Sensors Using the Same", filed on August 29, 2006; Publication No. WO07027691, on March 8, 2007. 101. U.S. Utility Patent Application Serial No. 14 / 419,375, titled "Computer Simulations for Testing and Monitoring Therapeutic Strategies for Stress Hyperglycemia", filed on February 3, 2015; U.S. Patent No. 10,438,700, issued on October 8, 2019. 102. International Patent Application Serial No. PCT / US2013 / 053664, titled "Computer Simulations for Testing and Monitoring Therapeutic Strategies for Stress Hyperglycemia", filed on August 5, 2013; Publication No. WO2014 / 022864, on February 6, 2014. 103. U.S. Utility Patent Application Serial No. 14 / 241,383, titled "Methods, Systems, and Computer-Readable Media for Adaptive Advisory Control of Diabetes", filed on February 26, 2014; U.S. Patent No. 11,024,429, issued on June 1, 2021. 104. International Patent Application Serial No. PCT / US2012 / 052422, titled "Methods, Systems, and Computer-Readable Media for Adaptive Advisory Control of Diabetes", filed on August 26, 2012; Publication No. WO2013 / 032965, on March 7, 2013. 105. U.S. Utility Patent Application Serial No. 14 / 128,811, titled "Methods and Apparatus for Modular Power Management and Protection of Critical Services in Ambulatory Medical Devices", filed on December 23, 2013; U.S. Patent No. 9,430,022, issued on August 30, 2016. 106. International Patent Application Serial No. PCT / US2012 / 043883, titled "Methods and Apparatus for Modular Power Management and Protection of Critical Services in Ambulatory Medical Devices", filed on June 22, 2012; Publication No. WO2012 / 178113, on December 27, 2012. 107. U.S. Utility Patent Application Serial No. 29 / 467,039, titled "Alarm Clock Display of Personal Blood Glucose Levels", filed on September 13, 2013. 108. International Patent Application Serial No. PCT / US2013 / 042745, titled "Insulin-Pramlintide Formulations and Methods of Making and Using the Same", filed on May 24, 2013; Publication No. WO2013 / 177565, on November 28, 2013. 109. U.S. Utility Patent Application Serial No. 13 / 637,359, titled "Methods, Systems, and Computer Program Products for Improving Glucose Sensor Accuracy Using Insulin Delivery Observations in Diabetes", filed on September 25, 2012; U.S. Patent No. 9,398,869, issued on July 26, 2016. 110. International Patent Application Serial No. PCT / US2011 / 029793, titled "Methods, Systems, and Computer Program Products for Improving Glucose Sensor Accuracy Using Insulin Delivery Observations in Diabetes", filed on March 24, 2011; Publication No. WO2011 / 119832, on September 29, 2011. 111. U.S. Utility Patent Application Serial No. 13 / 393,647, titled "Systems, Methods, and Computer Program Products for the Regulation of Insulin Delivery (AID) in Diabetes Using a Nominal Open Loop Profile", filed on March 1, 2012; Publication No. 2012 / 0245556, on September 27, 2012. 112. International Patent Application Serial No. PCT / US2010 / 047386, titled "Systems, Methods, and Computer Program Products for the Regulation of Insulin Delivery (AID) in Diabetes Using a Nominal Open Loop Profile", filed on August 31, 2010; Publication No. WO2011 / 028731, on March 10, 2011. 113. U.S. Utility Patent Application Serial No. 13 / 131,467, titled "Methods, Systems, and Computer Program Products for Tracking Glucose Variability in Diabetes," filed on May 26, 2011; U.S. Patent No. 9,317,657, issued on April 19, 2016. 114. International Patent Application Serial No. PCT / US2009 / 065725, titled "Methods, Systems, and Computer Program Products for Tracking Glucose Variability in Diabetes," filed on November 24, 2009; Publication No. WO2010 / 062898, on June 3, 2010. 115. U.S. Utility Patent Application Serial No. 12 / 975,580, titled "Methods, Systems, and Computer Program Products for Evaluating Glucose Control of Diabetes from Self-Monitoring Data," filed on December 22, 2010; Publication No. 2012 / 0004512, on January 5, 2012. 116. U.S. Utility Patent Application Serial No. 11 / 305,946, titled "Methods, Systems, and Computer Program Products for Evaluating Glucose Control of Diabetes from Self-Monitoring Data," filed on December 19, 2005; U.S. Patent No. 7,874,985, issued on January 25, 2011. 117. U.S. Utility Patent Application Serial No. 10 / 240,228, titled "Methods, Systems, and Computer Program Products for Evaluating Glucose Control of Diabetes from Self-Monitoring Data," filed on September 26, 2002; U.S. Patent No. 7,025,425, issued on April 11, 2006. 118. International Patent Application Serial No. PCT / US2001 / 009884, titled "Methods, Systems, and Computer Program Products for Evaluating Glucose Control in Diabetes," filed on March 29, 2001; Publication No. WO01 / 72208, on October 4, 2001. 119. U.S. Utility Patent Application Serial No. 12 / 664,444, titled "Methods, Systems, and Computer Simulation Environments for Testing Monitoring and Control Strategies in Diabetes," filed on December 14, 2009; U.S. Patent No. 10,546,659, issued on January 28, 2020. 120. International Patent Application Serial No. PCT / US2008 / 067725, titled "Methods, Systems, and Computer Simulation Environments for Testing Strategies for Monitoring and Control in Diabetes", filed on June 20, 2008; Publication No. WO2008 / 157781, on December 24, 2008. 121. U.S. Utility Patent Application Serial No. 12 / 516,044, titled "Methods, Systems, and Computer Program Products for Detecting Physical Activity Based on Heart Rate Changes in Diabetes, Evaluating Changes in Fasting Metabolic State, and Applying Closed and Open Control Loops", filed on May 22, 2009; U.S. Patent No. 8,585,593, issued on November 19, 2013. 122. International Patent Application Serial No. PCT / US2007 / 085588, titled "Methods, Systems, and Computer Program Products for Detecting Physical Activity Based on Heart Rate Changes in Diabetes, Evaluating Changes in Fasting Metabolic State, and Applying Closed and Open Control Loops", filed on November 27, 2007; Publication No. WO2008 / 067284, on June 5, 2008. 123. U.S. Utility Patent Application Serial No. 12 / 159,891, titled "Methods, Systems, and Computer Program Products for Evaluating Glucose Variability in Diabetes from Self-Monitoring Data", filed on July 2, 2008; U.S. Patent No. 11,355,238, issued on June 7, 2022. 124. International Patent Application Serial No. PCT / US2007 / 000370, titled "Methods, Systems, and Computer Program Products for Evaluating Glucose Variability in Diabetes from Self-Monitoring Data", filed on January 5, 2007; Publication No. WO07081853, on July 19, 2007. 125. U.S. Utility Patent Application Serial No. 11 / 943,226, titled "Systems, Methods, and Computer Program Code for Recognizing Patterns of Hyperglycemia and Hypoglycemia, Increasing Glucose Variability, and Ineffective Self-Monitoring in Diabetes", filed on November 20, 2007; Publication No. 2008 / 0154513, on June 26, 2008. 126. U.S. Utility Patent Application Serial No. 11 / 578,831, titled "Method, System, and Computer Program Product for Evaluating the Accuracy of a Glucose Monitoring Sensor / Device," filed on October 18, 2006; U.S. Patent No. 7,815,569, issued on October 19, 2010. 127. International Patent Application Serial No. US2005 / 013792, titled "Method, System, and Computer Program Product for Evaluating the Accuracy of a Glucose Monitoring Sensor / Device," filed on April 21, 2005; Publication No. WO05106017, on November 10, 2005. 128. U.S. Utility Patent Application Serial No. 10 / 524,094, titled "Method, System, and Computer Program Product for Processing Self-Monitoring Blood Glucose (SMBG) Data to Expand Diabetes Self-Management," filed on February 9, 2005; U.S. Patent No. 8,538,703, issued on September 17, 2013.

[0236] 129. International Patent Application Serial No. PCT / US2003 / 025053, titled "Management and Processing of Self-Monitoring Blood Glucose," filed on August 8, 2003; Publication No. WO2004 / 015539, on February 19, 2004. 130. International Patent Application Serial No. PCT / US2002 / 005676, titled "Method and Apparatus for Early Diagnosis of Subacute Potentially Disastrous Diseases," filed on February 27, 2002; Publication No. WO02 / 67776, on September 6, 2002. 131. U.S. Utility Patent Application Serial No. 09 / 793,653, titled "Method and Apparatus for Early Diagnosis of Subacute Potentially Disastrous Diseases," filed on February 27, 2001; U.S. Patent No. 6,804,551, issued on October 12, 2004. 132. U.S. Utility Patent Application Serial No. 10 / 069,674, titled "Method and Apparatus for Predicting the Risk of Hypoglycemia," filed on February 22, 2002; U.S. Patent No. 6,923,763, issued on August 2, 2005. 133. International Patent Application Serial No. US00 / 22886, titled "Method and Apparatus for Predicting the Risk of Hypoglycemia," filed on August 21, 2000; Publication No. WO01 / 13786, on March 1, 2001.

Description of Symbols

[0237] 100 System 102 Data Store for Clustered Data 103 Data Store for Classified Blood Glucose Measurement Values 104 Processor 106 Memory 108 Instruction 110 Other Devices 112 Data Source

Claims

1. A system for processing glucose data by efficient glucose database management, comprising: A physical data store including glucose measurement data and a representation for at least one cluster of the glucose measurement data, the representation approximating a blood glucose profile vector array for one cluster of a plurality of glucose profiles segmented by a plurality of time ranges; a physical data store; A processor; and a computer memory configured by stored instructions; Comprising: When the instructions are executed, Receiving a glucose measurement value; Converting the glucose measurement value into a vector format; Searching the physical data store by comparing a newly received glucose measurement value with the centroid of a cluster using a similarity metric; Classifying the newly received glucose measurement value by a cluster having a matching similarity metric based on the comparison; Attributing a treatment to the newly received glucose measurement value; A system that causes the processor to perform the above.

2. The instructions are: Storing the classification of the newly received glucose measurement value in a data store that communicates with one or more of a prediction modeling system, a decision support system, an insulin delivery system, an insulin monitoring system, or an automatic control system configured to use the classification as an input; Transmitting the classification of the newly received glucose measurement value to one or more of a prediction modeling system, a decision support system, an insulin delivery system, an insulin monitoring system, or an automatic control system configured to use the classification as an input; or Using the classification of the newly received glucose measurement value to monitor, analyze, or manipulate the concentration of glucose levels in a body fluid. The system according to claim 1, which causes the processor to perform one or more of the above.

3. The system according to claim 1, wherein the instructions cause the processor to receive the glucose measurement value from a glucose measurement device.

4. The system according to claim 3, comprising the glucose measurement device.

5. The data store that communicates with one or more of the prediction modeling system, the decision support system, the insulin delivery system, the insulin monitoring system, or the automatic control system; One or more of the prediction modeling system, the decision support system, the insulin delivery system, the insulin monitoring system, or the automatic control system; The system according to claim 2, comprising:

6. The instructions cause The system according to claim 1, wherein the processor is caused to calculate the Euclidean distance between one or more newly received glucose measurement values and one or more centroids as the similarity measure.

7. Comprising a plurality of clusters, The plurality of clusters are Generating an array of glucose measurement values for each time range, wherein the plurality of arrays form a blood glucose profile vector; Assigning weights to the arrays; Applying an iterative hierarchical clustering technique that varies the weights until one or more clusters are generated that approximate one or more of the blood glucose profile vectors; Defining one of the plurality of clusters by the centroid of the cluster; The system according to claim 1, generated by:

8. The iterative hierarchical clustering technique computes the R value by linear regression on the array and changes the weighting to maximize the R value. The system according to claim 7. 2 2 ​​

9. The system according to claim 1, wherein the plurality of time ranges include five time ranges.

10. The plurality of time ranges are Level 2 hypoglycemia below glucose measurement - 1; Level 1 hypoglycemia within the range of glucose measurement - 2 to glucose measurement - 3; Target range (TIR) within the range of glucose measurement - 4 to glucose measurement - 5; Level 1 hyperglycemia within the range of glucose measurement - 6 to glucose measurement - 7; Level 2 hyperglycemia above glucose measurement - 8, The system according to claim 1, comprising:

11. Glucose measurement - 1 is 54 mg / dl, Glucose measurement - 2 is 54 mg / dl, Glucose measurement - 3 is 70 mg / dL, Glucose measurement - 4 is 70 mg / dL, Glucose measurement - 5 is 180 mg / dL, Glucose measurement - 6 is 180 mg / dL, Glucose measurement - 7 is 250 mg / dL, Glucose measurement - 8 is 250 mg / dL, The system according to claim 10.

12. The glucose measurement values include a plurality of glucose profiles of an individual, each glucose profile includes a plurality of glucose measurement values acquired over a predetermined period, and the instructions cause the processor to compile the plurality of glucose profiles of the individual into a time series of single glucose measurement values, classify one or more glucose profiles using one or more clusters, and generate a sequence of indices representing the classification of the one or more glucose profiles in the single glucose measurement value time series, The system according to claim 1, which causes the processor to perform the above.

13. The instructions cause the processor to generate a trace representing the glycemic variability of the individual using the sequence of indices, for the system according to claim 12.

14. The instructions cause the processor to generate an approximated ambulatory glucose profile (AGP) using the sequence of indices, for the system according to claim 12.

15. One or more of the plurality of glucose profiles include a continuous glucose monitoring (CGM) profile including glucose measurement values acquired over 24 hours, for the system according to claim 1.

16. One or more of the plurality of glucose profiles of the individual include a continuous glucose monitoring (CGM) profile including glucose measurement values acquired over 24 hours, for the system according to claim 12.

17. A method for processing glucose data for efficient glucose database management, receiving glucose measurement values, converting the glucose measurement values into vector form, searching a physical data store by comparing newly received glucose measurement values with the centroid of a cluster using a similarity measure, wherein the physical data store includes glucose measurement value data and a representation for at least one cluster of the glucose measurement value data, and the representation approximates a glucose profile vector for one cluster of a plurality of glucose profiles segmented by a plurality of time ranges. classifying the newly received glucose measurement by a cluster having a similarity measure matched based on the comparison; attributing a treatment to the newly received glucose measurement; A method comprising: **Claim 18** The method according to claim 17, comprising calculating a Euclidean distance between one or more of the newly received glucose measurements and one or more of the centroids as the similarity measure. **Claim 19** The physical data store generates an array of glucose measurements for each time range, with multiple arrays forming a blood glucose profile vector; assigning weights to the arrays; applying an iterative hierarchical clustering technique that varies the weights until one or more clusters are generated that approximate one or more blood glucose profile vectors; defining one of the clusters of the cluster set by the centroid of the cluster; The method according to claim 16, comprising a plurality of clusters generated by: **Claim 20** Through the iterative hierarchical clustering technique, calculate the R value by linear regression of the array, and include the step of changing the weighting to maximize the R value. The method according to claim 19. 2 value by computer calculation, and the R 2 value. The method according to claim 19, comprising the step of changing the weighting to maximize the

Images