Summarizing data from continuous glucose monitors using the cgmstats command

Abstract

In this article, we present the cgmstats command, which analyzes continuous glucose monitoring (CGM) data. The use of wearable CGMs is growing rapidly. The latest generation of CGM systems does not require fingerstick calibration, is minimally invasive, and is frequently used in research studies. CGM sensors are typically worn for up to 2 weeks and record interstitial glucose measurements every minute to every 15 minutes, depending on the sensor used. CGM systems generate hundreds of measurements per day and thousands of measurements in one person over a single wear. There is a need for tools that allow researchers to efficiently organize and summarize the wealth of data on glucose patterns produced by CGM systems. The cgmstats command generates CGM summary measures for data from various CGM systems and allows the user to flexibly define ranges and generate data visualizations. In this article, we provide an overview of the cgmstats command and examples of its use. The cgmstats command supports rigorous and reproducible analyses of CGM data.

Keywords

st0804 cgmstats continuous glucose monitoring metrics glucose profile

1 Introduction

Continuous glucose monitoring systems (CGMs) were originally designed for the clinical management of type 1 diabetes. However, their use has expanded rapidly, and they are now used by patients with type 2 diabetes and even individuals without diabetes. There have been numerous randomized clinical trials demonstrating the benefits of CGM use for managing glucose control (Beck et al. 2017; Martens et al. 2021), and CGMs have been incorporated into large, population-based studies to examine glucose patterns in the general community and their associations with clinical outcomes (Daya et al. 2024; Shah et al. 2019; Selvin et al. 2021; Spartano et al. 2025). CGM systems generate a large amount of data (that is, ∼1,300 to ∼21,000 readings over 15 days, depending on the sensor) in different formats. Flexible and reproducible tools are needed for researchers to analyze and interpret these data. While CGM manufacturers produce CGM summary reports for patients and their providers, these reports are designed for clinical use and typically summarize up to ∼2 weeks of data on a single person with a limited number of CGM summary statistics.

There is currently no standardized approach to the processing or analysis of CGM data. This lack of standardization in the field is complicated by the diversity of data output formatting across CGM systems, making it difficult to directly compare results including CGM summary statistics across studies. Thus, there is a need for flexible statistical commands that can process a wide variety of data and conduct analyses using different approaches. For example, these commands should be able to vary cutpoints for time in range and time above and below range, alter the definition of hypoglycemic or hyperglycemic episodes based on level and duration, require a certain percentage of valid readings (for example, 70%) for the summary metrics to be considered reliable, and be able to drop the first or last 24 hours of data. Typically, CGM sensor data are considered less reliable during the first 24 hours of wear. During this time, a CGM sensor may have unexpectedly high or low readings as it is initializing and calibrating. Over time, sensor degradation or changes in adhesion to the skin may lead to inaccurate readings or signal interruptions toward the end of the wear period. There also may be gaps in the data because of issues with the sensor or transmitter (that is, sensor displacement, poor adhesion, software glitches, etc.). There is no standard approach for handling these issues.

In prior studies, including major randomized clinical trials, it is unclear if some CGM data points were excluded or if all available sensor readings were used (Aleppo et al. 2017). For example, some studies drop the first and last days of CGM data, but others do not (Beck et al. 2017; Wan et al. 2018). The handling of missing data is also variable. While some studies impute small gaps (that is, less than 30 minutes) using interpolation methods, others do not (Kuang et al. Forthcoming). Large gaps may lead to exclusion of participants if the missing data exceed a predefined threshold to deem the summary statistic reliable (that is, the guidelines suggest that a minimum of 70% of data are captured from the total CGM wear time).

The lack of standardization in data processing methods across studies including clinical trials involving CGM data suggests a need for clearer reporting and methodological consistency to enhance the reliability of findings.

Automating the calculation of CGM metrics is efficient and ensures an accurate and reproducible approach. Several commands exist, primarily in R or as web-based or desktop applications, to process and summarize CGM data (Karakus, Snell-Bergeon, and Akturk 2025). To our knowledge, there are no commands that provide a flexible and automated approach to process and comprehensively analyze CGM data. Also, existing commands often lack the flexibility to accommodate data from the more than 10 major CGM systems used globally that store glucose readings at various intervals. They also lack the flexibility to allow users to easily drop a specified number of 24-hour periods from the start or end of the wear period or restrict the analysis to a specified period according to dates and times, all in one command. Most software commands do not automatically generate individual time-series data for a given person to examine how an individual’s CGM summary metrics vary by day.

We created the first command to process, summarize, and visualize data from CGM systems. cgmstats processes data obtained from different CGM systems containing three columns of data (ID, date or timestamp, and CGM glucose reading) and outputs a dataset with various CGM summary metrics, including standard clinical summary measures used in the management of diabetes. These include percent time in range, time spent at high (hyperglycemia) or low levels of glucose (hypoglycemia), the count and duration of hypoglycemic and hyperglycemic episodes, and measures of glycemic variability (Battelino et al. 2023; American Diabetes Association Professional Practice Committee 2025).

The cgmstats command provides user flexibility to analyze CGM data from various sensors and different generations of devices from the same manufacturers, to define thresholds and duration of hypoglycemic and hyperglycemic episodes, and to output CGM summary metrics into a new dataset by person or by day or a time-series dataset by person and day. The command accepts CGM data with gaps. It also lets users visualize CGM glucose tracings over the wear period with the option to fill in gaps in the data using linear interpolation.

Because CGMs are increasingly used by patients and in research studies, there is a growing need for software to summarize and visualize CGM time series using a flexible approach and with automated output of standard CGM metrics that are consistent with clinical practice. The cgmstats command fulfills this need in Stata.

2 The cgmstats command

2.1 Description

The cgmstats command processes the CGM data according to the user’s specifications and creates a dataset with summary CGM metrics. By default, cgmstats generates measures using definitions recommended in clinical guidelines (Battelino et al. 2023). Users also have the option to create CGM metrics using custom cutpoints. The command allows users to define the period of analysis and includes functions to exclude biologically implausible values (for example, low values caused by pressure placed on the sensor during sleep; high or low oxygen tensions; improper sensor calibration). The summary CGM metrics are calculated overall and separately by daytime and nighttime. The command accepts glucose measurements in units of mg/dL (default) or mmol/L and outputs metrics in the same unit.

The cgmstats command produces and saves histograms of user-specified CGM summary metrics. It also plots CGM tracing per person if specified by the user.

2.2 Syntax

The input dataset or datasets must contain three columns with the following information: ID (unique identifier for each person); timestamp (date and time of glucose reading); and sensor glucose reading. The user specifies the variable name for each of the three columns.

Raw CGM data are typically received in the comma-separated values format. The user may prepare the respective .dta files using the command import excel.

All input datasets should be saved in the user-specified directory or, if not specified in the command syntax, the working directory (by default).

The command syntax is

cgmstats, id(varname) glucose(varname) time(varname) dtadir(string) [freq(#) unit(mmol/L) by(day | id day) keep(stri'ng) hyper(numlist) hypo(numlist) hyper_exc_lngth(#) hypo_exc_lngth(#) daystart(string) dayend(string) firsthours(#) lasthours(#) timebefore(timestamp) timeafter(timestamp) auc hist(varlist[, freqbw(#)]) plot(numlist[, nostats fill]) savecombdta(string) savecombdir(string) savesumdta(string) savesumdir(string) saveplotdir(string) ]

If timestamp is in the incorrect format, the user will receive an error that says, for example,

Additionally, if duplicate timestamps are found, the user will receive the following error:

If the directory containing the input .dta files is not specified, the user will receive the following error:

2.3 Options

id(varname) is the unique identifier (numeric or string variable) of the patient or participant. id() is required.

glucose(varname) is the numeric CGM glucose values. glucose() is required.

time(varname) is the timestamp of CGM measurement. The timestamp must be in the format %tcCCYY/NN/DD_HH:MM (for example, 2014/12/01 01:30). time() is required.

dtadir(string) specifies the directory containing all the input files in .dta format. dtadir() is required.

freq(#) establishes CGM recording frequency (in minutes). The default is freq(15).

unit(mmol/L) specifies that glucose units be in mmol/L. The default is mg/dL.

by(day) or by(id day) produces CGM metric summary files (datasets and plots) by day of CGM wear (that is, by(day)) or produces one CGM metric summary file by ID and day of wear (multiple rows per ID) (that is, by(id day)).

keep(string) contains conditional statements to subset data according to glucose values or ID (that is, keep(if sensorglucose>50)).

hyper(numlist) defines glucose value cutpoints for a hyperglycemic episode. An episode is defined as no values below glucose value cutpoints for a certain period of time. Default values are 140 mg/dL (7.8 mmol/L), 180 mg/dL (10 mmol/L), and 250 mg/dL (13.9 mmol/L). If unit(mmol/L) is specified, then cutpoints should be in units of mmol/L. Decimals are accepted, but decimal points will be converted to _ in variable names.

hypo(numlist) defines glucose value cutpoints for a hypoglycemic episode. An episode is defined as no values above glucose value cutpoints for a certain period of time. Default values are 54 mg/dL (3 mmol/L) and 70 mg/dL (3.9 mmol/L). If unit(mmol/L) is specified, then cutpoints should be in units of mmol/L. Decimals are accepted, but decimal points will be converted to _ in variable names.

hyper_exc_lngth(#) defines length (in minutes) of a hyperglycemic episode. The default is hyper_exc_lngth(30). During this period, there are no values below X (as defined by the option hyper()).

hypo_exc_lngth(#) defines length (in minutes) of a hypoglycemic episode. The default is hypo_exc_lngth(30). During this period, there are no values above X (as defined by the option hypo()).

daystart(string) defines the time indicating the start of the day (HH:MM, 24-hour scale). The default is daystart(06:00).

dayend(string) defines the time indicating the end of the day (HH:MM, 24-hour scale). The default is dayend(22:00).

firsthours(#) defines the number of hours to drop from the start of the CGM wear period. This option accepts decimals (that is, 0.5 for 30 minutes).

lasthours(#) defines the number of hours to drop from the end of the CGM wear period. This option accepts decimals (that is, 0.5 for 30 minutes).

timebefore(timestamp) subsets data to time before the specified timestamp, that is, timebefore(”2021/09/01 12:00”).

timeafter(timestamp) subsets data to time after the specified timestamp, that is, timeafter(”2023/06/01 12:00”).

auc calculates the area under the curve over glucose value cutpoints defined by the option hyper().

hist(varlist[, freq bw(#) ]) plots histograms of user-specified CGM summary statistics overlaid with kernel density plots (that is, hist(mean_sensor cv_sensor)) by default. If option freq is specified, the frequency histograms are plotted instead without kernel density plots (that is, hist(mean_sensor, freq)). bw() alters the bar widths of the histogram.

plot(numlist[, nostats fill ]) plots CGM glucose tracings over the entire CGM wear period with key CGM summary metrics (this can be turned off using option nostats, that is, plot(4 10, nostats)). Users define the two cutpoints (low and high) for time in range, and these cutpoints must be contained in options hypo() and hyper(). If a decimal was specified in option hypo() or hyper(), it must be rounded to the nearest whole number for the plot() option. fill fills in gaps in CGM glucose tracings using linear interpolation.

savecombdta(string) specifies the user-specified name of the .dta file containing the combined input .dta files.

savecombdir(string) specifies the user-specified file path where the combined input .dta file (named in the savecombdta() option) is saved. If the directory is not specified along with savecombdta(), the combined input .dta file will be saved in the user’s working directory by default.

savesumdta(string) specifies the user-specified name of the .dta file containing the summary CGM metrics.

savesumdir(string) specifies the user-specified file path where the summary CGM metric dataset (named in the savesumdta() option) is saved. If the directory is not specified along with savesumdta(), the summary CGM metric dataset will be saved in the user’s working directory by default.

saveplotdir(string) specifies the user-specified file path where any plots (histogram as specified in the hist() option and CGM glucose tracings as specified in the plot() option) are saved. If saveplotdir() is not specified, then plots are not saved.

3 Output

After users run the cgmstats command, which contains the variable names for ID, timestamp, and glucose measurement at minimum, the result is a dataset containing the following CGM metrics (output variables):

If X_i, contains a decimal point, it is converted to an underscore in the variable name. For example, if X_i, is 2.9, the variable name would be percent_time_under_2_9.

The output dataset is saved in the specified working directory ($dir) and titled with the option savesumdta(). When by(day) is used, the output dataset is titled with savesumdta() and the day number; for example, savesumdta(cgm_summary_file) would create cgm_summary_file1 for day 1, cgm_summary_file2 for day 2, and so on.

4 Example 1

This example (using cgm5ex1.dta and cgm5ex2.dta) inputs a CGM dataset from a system that recorded glucose every 5 minutes (versus the default of 15 minutes), denoted using the option freq(5). Here the user includes CGM readings occurring any time before 12 p.m. on 2/6/2022 (timebefore(”2022/02/06 12:00”)). The example also specifies that the CGM readings were measured in mmol/L (the option unit(mmol/L)) and the summary statistics outputted in mmol/L.

Note that the decimal point in “3.9” (one of the default values for hypo() when units are mmol/L and hypo() values are not specified) is converted to an underscore in the output variable name, percent_time_3_9_10. Stata does not accept decimal points in variable names.

This is a snapshot of the output dataset:

Figure 1 displays frequency histograms of the CGM summary statistics that we specified in the hist() option of the previous cgmstats command.

Figure 1.

Frequency histograms of CGM specified summary statistics

The CGM glucose tracings for each individual in the dataset are output using the option plot(3.9 10). Figures 2 and 3 are the tracings for the first two individuals in the dataset:

Figure 2.

Individual CGM glucose tracing with summary data displayed, example 1, ID 1

Figure 3.

Individual CGM glucose tracing with summary data displayed, example 1, ID 2

The lighter shading indicates time in range, and the darker shading indicates time below range, as specified by the user or according to the default (time in range 70-140 mg/dL or 3.9-7.8 mmol/L).

In the example CGM tracings above, there are gaps between CGM glucose readings. If we specify the option fill, these gaps are filled using linear interpolation:

In both versions, the summary statistics printed on the plots are from the actual data (no interpolation). Summary statistics can be suppressed using the option plot(3.9 10, nostats).

5 Example 2

This example (using cgm15ex1.dta and cgm15ex2.dta) specifies that CGM summary metrics are outputted in one dataset by ID and by day (that is, multiple rows per person) using the option by(id day). In this example, we drop the first hour of CGM wear (firsthours(1)) and define hypoglycemic episodes as 15 minutes below 54 or 70 mg/dL and hyperglycemic episodes as 45 minutes above 140, 180, or 250 mg/dL.

This is a snapshot of the output dataset:

The CGM glucose tracings for each person and each day of wear are outputted using the option plot(70 140). Figures 6 and 7 are examples, showing the tracings for day 5 and 11 of an individual (subjectid = 6) in the dataset:

Figure 4.

Individual CGM glucose tracing with summary data displayed and gaps filled using linear interpolation, example 1, ID 1

Figure 5.

Individual CGM glucose tracing with summary data displayed and gaps filled using linear interpolation, example 1, ID 2

Figure 6.

Individual CGM tracing by day with day-specific summary statistics, example 2, ID 6, day 5

Figure 7.

Individual CGM tracing by day with day-specific summary statistics, example 2, ID 6, day 11

6 Conclusions

This article introduced cgmstats, a command that provides a flexible framework for processing CGM data and generating summaries and visualizations of data from several CGM systems, which are increasingly available and used in research studies and clinical trials. By streamlining workflows, cgmstats enhances rigor and reproducibility in CGM research and lowers barriers to incorporating CGM data into epidemiological, clinical, and translational research.

Emerging research using CGM data involves advanced approaches including machine learning models (Dave et al. 2020; Klonoff et al. 2025). Although the cgmstats command focuses on descriptive statistics and visualizations, these are a critical first step for more complex analyses, and cgmstats can provide the input required for machine learning models and other advanced modeling techniques. Integrating CGM data with information from other wearable devices or applications, such as dietary logs and physical activity trackers, may also be of interest to researchers and can be facilitated by the output produced from this command.

The functions and flexibility offered in cgmstats provide Stata users with an efficient and automated method to process, summarize, and visualize CGM data from multiple systems. This cgmstats command is well positioned to evolve alongside advances in digital health research, supporting broader applications and accelerating the next generation of high-quality scientific research using these clinically important data.

Supplemental Material

sj-txt-1-stj-10.1177_1536867X261450271 - Supplemental material for Summarizing data from continuous glucose monitors using the cgmstats command

Supplemental material, sj-txt-1-stj-10.1177_1536867X261450271 for Summarizing data from continuous glucose monitors using the cgmstats command by Natalie Daya Malek, Dan Wang, Sui Zhang, Michael Fang, Amelia Wallace, Scott Zeger and Elizabeth Selvin in The Stata Journal

Supplemental Material

sj-dta-4-stj-10.1177_1536867X261450271 - Supplemental material for Summarizing data from continuous glucose monitors using the cgmstats command

Supplemental material, sj-dta-4-stj-10.1177_1536867X261450271 for Summarizing data from continuous glucose monitors using the cgmstats command by Natalie Daya Malek, Dan Wang, Sui Zhang, Michael Fang, Amelia Wallace, Scott Zeger and Elizabeth Selvin in The Stata Journal

Supplemental Material

sj-dta-3-stj-10.1177_1536867X261450271 - Supplemental material for Summarizing data from continuous glucose monitors using the cgmstats command

Supplemental material, sj-dta-3-stj-10.1177_1536867X261450271 for Summarizing data from continuous glucose monitors using the cgmstats command by Natalie Daya Malek, Dan Wang, Sui Zhang, Michael Fang, Amelia Wallace, Scott Zeger and Elizabeth Selvin in The Stata Journal

Supplemental Material

sj-dta-2-stj-10.1177_1536867X261450271 - Supplemental material for Summarizing data from continuous glucose monitors using the cgmstats command

Supplemental material, sj-dta-2-stj-10.1177_1536867X261450271 for Summarizing data from continuous glucose monitors using the cgmstats command by Natalie Daya Malek, Dan Wang, Sui Zhang, Michael Fang, Amelia Wallace, Scott Zeger and Elizabeth Selvin in The Stata Journal

Supplemental Material

sj-dta-1-stj-10.1177_1536867X261450271 - Supplemental material for Summarizing data from continuous glucose monitors using the cgmstats command

Supplemental material, sj-dta-1-stj-10.1177_1536867X261450271 for Summarizing data from continuous glucose monitors using the cgmstats command by Natalie Daya Malek, Dan Wang, Sui Zhang, Michael Fang, Amelia Wallace, Scott Zeger and Elizabeth Selvin in The Stata Journal

Footnotes

7 Acknowledgments

We thank the members of Dr. Elizabeth Selvin’s Diabetes Data group for their invaluable support with this work, including pilot testing and feedback. Dr. Selvin is supported by National Institutes of Health and National Institute of Diabetes and Digestive and Kidney Diseases grant R01 DK128837 and a Merit Award from the American Heart Association. Dr. Fang is supported by National Institutes of Health and National Institute of Diabetes and Digestive and Kidney Diseases career development award K01 DK138273.

8 Programs and supplemental materials

To install the software files as they existed at the time of publication of this article, type

About the authors

Natalie Daya Malek is a research associate at the Johns Hopkins Bloomberg School of Public Health (Welch Center for Prevention, Epidemiology, and Clinical Research).

Dan Wang and Sui Zhang are senior biostatisticians at the Johns Hopkins Bloomberg School of Public Health (Welch Center for Prevention, Epidemiology, and Clinical Research).

Michael Fang is an assistant professor at the Johns Hopkins Bloomberg School of Public Health (Welch Center for Prevention, Epidemiology, and Clinical Research).

Amelia Wallace is an assistant scientist at the Johns Hopkins Bloomberg School of Public Health (Welch Center for Prevention, Epidemiology, and Clinical Research).

Scott Zeger is a professor at Johns Hopkins Bloomberg School of Public Health in the Department of Biostatistics.

Elizabeth Selvin is a professor at Johns Hopkins Bloomberg School of Public Health and the director of the Welch Center for Prevention, Epidemiology, and Clinical Research.

All authors are longtime Stata users.

References

Aleppo

Ruedy

K. J.

Riddlesworth

T. D.

Kruger

D. F.

Peters

A. L.

Hirsch

Bergenstal

R. M.

Toschi

Ahmann

A. J.

Shah

V. N.

Rickels

M. R.

Bode

B. W.

Philis-Tsimikas

Pop-Busui

Rodriguez

Eyth

Bhargava

Kollman

Beck

R. W.

. 2017. 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 40: 538–545. 10.2337/dc16-2482.

American Diabetes Association Professional Practice Committee. 2025. Diabetes technology: Standards of care in diabetes. Diabetes Care 48: S146–S166. 10.2337/dc25-S007.

Battelino

Alexander

C. M.

Amiel

S. A.

Arreaza-Rubin

Beck

R. W.

Bergenstal

R. M.

Buckingham

B. A.

Carroll

Ceriello

Chow

Choudhary

Danne

Dutta

Gabbay

Garg

Heverly

Hirsch

I. B.

Kader

Kenney

, et al. 2023. Continuous glucose monitoring and metrics for clinical trials: An international consensus statement. Lancet Diabetes Endocrinol 11: 42–57. 10.1001/jama.2016.19975.

Beck

R. W.

Riddlesworth

Ruedy

Ahmann

Bergenstal

Kollman

Haller

Kruger

McGill

J. B.

Polonsky

Toschi

Wolpert

Price

. 2017. 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 317: 371–378. 10.1001/jama.2016.19975.

Dave

DeSalvo

D. J.

Haridas

McKay

Shenoy

Koh

C. J.

Lawley

Erraguntla

. 2020. Feature-based machine learning model for real-time hypoglycemia prediction. Journal of Diabetes Science and Technology 15: 842–855. 10.1177/1932296820922622.

Daya

N. R.

Fang

Wang

Valint

Windham

B. G.

Coresh

Echouffo- Tcheugui

J. B.

Selvin

. 2024. Glucose abnormalities detected by continuous glucose monitoring in very old adults with and without diabetes. Diabetes Care 48: 416–421. 10.2337/dc24-1990.

Karakus

K. E.

Snell-Bergeon

J. K.

Akturk

H. K.

. 2025. Comparison of computational statistical packages for the analysis of continuous glucose monitoring data with a reference software, “Ambulatory glucose profile,” in type 1 diabetes. Diabetes Technology and Therapeutics 27: 202–208. 10.1089/dia.2024.0410.

Klonoff

D. C.

Bergenstal

R. M.

Cengiz

Clements

M. A.

Espes

Espinoza

Kerr

Kovatchev

Maahs

D. M.

Mader

J. K.

Mathioudakis

Metwally

A. A.

Shah

S. N.

Sheng

Snyder

M. P.

Umpierrez

Shao

M. M.

Scheideman

A. F.

Ayers

A. T.

C. N.

Healey

. 2025. CGM data analysis 2.0: Functional data pattern recognition and artificial intelligence applications. Journal of Diabetes Science and Technology 19: 1515–1527. 10.1177/1932296825135322.

Kuang

Siddique

Shcholtens

. Forthcoming. Imputation of missing continuous glucose monitor data. Journal of Diabetes Science and Technology, 10.1177/19322968241308217.

10.

Martens

Beck

R. W.

Bailey

Ruedy

K. J.

Calhoun

Peters

A. L.

Pop- Busui

Philis-Tsimikas

Bao

Umpierrez

Davis

Kruger

Bhargava

Young

McGill

J. B.

Aleppo

Nguyen

Q. T.

Orozco

Biggs

Lucas

K. J.

, et al. 2021. Effect of continuous glucose monitoring on glycemic control in patients with type 2 diabetes treated with basal insulin: A randomized clinical trial. Journal of the American Medical Association 325: 2262–2272. 10.1001/jama.2021.7444.

11.

Selvin

Wang

Tang

Minotti

Echouffo-Tcheugui

J. B.

Coresh

. 2021. Glucose patterns in very old adults: A pilot study in a community-based population. Diabetes Technology and Therapeutics 23: 737–744. 10.1089/dia.2021.0156.

12.

Shah

V. N.

DuBose

S. N.

Beck

R. W.

Peters

A. L.

Weinstock

R. S.

Kruger

Tansey

Sparling

Woerner

Vendrame

Bergenstal

Tamborlane

W. V.

Watson

S. E.

Sherr

. 2019. Continuous glucose monitoring profiles in healthy nondiabetic participants: A multicenter prospective study. Journal of Clinical Endocrinology and Metabolism 104: 4356–4364. 10.1210/jc.2018-02763.

13.

Spartano

N. L.

Sultana

Lin

Cheng

Fei

Murabito

J. M.

Walker

M. E.

Wolpert

H. A.

Steenkamp

D. W.

. 2025. Defining continuous glucose monitor time in range in a large, community-based cohort without diabetes. Journal of Clinical Endocrinology and Metabolism 110: 1128–1134. 10.1210/clinem/dgae626.

14.

Wan

Reza Skandari

Minc

Nathan

A. G.

Zarei

Winn

A. N.

O’Grady

Huang

E. S.

. 2018. Cost-effectiveness of initiating an insulin pump in T1D adults using continuous glucose monitoring compared with multiple daily insulin injections: The DIAMOND Randomized Trial. Medical Decision Making 38: 942–953. 10.1177/0272989X18803109.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

1.95 MB

2.05 MB

4.39 MB

0.00 MB

4.77 MB

0.00 MB