OBJECTIVE

We captured continuous glucose monitoring (CGM) metrics from a large online survey of adults with type 1 diabetes to determine how glycemic outcomes varied by insulin delivery form.

RESEARCH DESIGN AND METHODS

Adults with type 1 diabetes from the T1D Exchange Registry/online communities completed the survey and contributed retrospective CGM data for up to 1 year. Self-reported glycemic outcomes and CGM measures were described overall and by insulin delivery method.

RESULTS

The 926 participants completed the survey and provided CGM data. Mean ± SD age was 41.9 ± 15.7 years, and 50.8% reported using automated insulin delivery (AID). While AID users spent more time in range, 27.9% did not achieve time in range targets, 15.5% reported severe hypoglycemic events (SHEs), and 16.0% had CGM-detected level 2 hypoglycemic events.

CONCLUSIONS

Despite use of diabetes technologies, many individuals are unable to achieve glycemic targets and experience severe hypoglycemia, highlighting the need for novel treatments.

Despite advances in type 1 diabetes management (i.e., continuous glucose monitoring [CGM], automated insulin delivery [AID] systems), many individuals do not meet hemoglobin A1c (HbA1c) targets and experience severe hypoglycemic events (SHEs) (1). With widespread CGM uptake, detailed metrics are available to provide insight into glucose levels, which can guide clinical evaluation and management (2,3). While HbA1c and SHE assessments are part of routine clinical care, their comparisons with CGM-derived metrics, like time spent with glucose in target ranges and sensor-detected hypoglycemia, are not well characterized across insulin delivery methods. Previously, we reported survey results of 2,044 individuals with type 1 diabetes showing that a high proportion did not achieve glycemic targets, experienced SHEs, and had impaired awareness of hypoglycemia (IAH), despite diabetes technology use (1). A subset of this previous survey provided actual CGM data, which allows us to assess concordance among patient-reported glycemic control measures, hypoglycemia, and CGM-derived metrics and identify the benefits of current AID systems in overall glycemic management and gaps in care.

An observational, retrospective study with cross-sectional elements was conducted February–April 2021 in adults ≥18 years old with type 1 diabetes for ≥2 years. Participants enrolled from the T1D Exchange Registry/online communities. The full survey sample, with or without CGM data, has previously been described (1). Here, the data reflect participants who contributed retrospective CGM data for ≤1 year (CGM subset). Qualifying for this subset requires that CGM data be included for ≥14 consecutive days, and CGM had to be used ≥70% of the time.

Survey participants self-reported their most recent HbA1c and number of SHEs (defined as severe low blood glucose and needing help from another person for treatment) over the past 12 months. Participants also completed the modified Gold questionnaire; participants scoring ≥4 were considered to have IAH (4) (Supplementary Table 1).

All CGM data from the preceding 12 months were used to calculate glucose management indicator (GMI), time in CGM ranges, coefficient of variation (CV), and level 2 sensor-detected prolonged hypoglycemia (SDPH) events. The GMI target was <7%; the CV target was <36%. Amount of time spent in various glycemic ranges was calculated according to consensus guidelines (5): very high, >250 mg/dL, target <5% of time; time above range (TAR), >180 mg/dL, target <25%; time in range (TIR), 70–180 mg/dL, target >70%; time in tight range (TITR), 70–140 mg/dL, current suggested target >50%; time below range (TBR), <70 mg/dL, target <4%; and very low, <54 mg/dL, target <1%. The proportion of participants achieving times in target range (70–180 mg/dL) and a composite measure of achieving TIR and TBR were calculated. SDPH was defined according to initial CGM glucose <54 mg/dL for ≥120 min and ending with glucose ≥70 mg/dL for ≥15 min or records not available for ≥120 min following CGM <54 mg/dL.

Analyses were descriptive with summary statistics (i.e., means/SDs for continuous variables and counts/proportions for categorical variables) and 95% CI. Concordance between self-reported and CGM-derived measures was assessed with use of contingency tables and Spearman correlation coefficient. Summary data are reported overall and by insulin delivery method: pump therapy (CGM+pump group), AID systems (CGM+AID), or multiple daily injections (MDI) (CGM+MDI).

Data and Resource Availability

Summary data may be made available on reasonable request via e-mail to the lead and corresponding authors.

A total of 926 participants contributed CGM data for ≤1 year, answered one or more diabetes survey questions, and provided CGM use status (Supplementary Fig. 1). Full survey results have previously been reported (1).

Among participants (72.8% female, 95.7% White) mean ± SD age was 41.9 ± 15.7 years and type 1 diabetes duration 25.0 ± 15.5 years (Table 1). Insulin delivery methods were as follows: MDI used by 17.5%, pump 31.6%, and AID system 50.8%. Self-reported CGM wear frequency was reported as 100% by 93.6% participants, and 41.1% participants used CGM for at least 5 years. Mean duration of available CGM data was 41.6 ± 14.9 weeks (range 1.6–52.1). Baseline characteristics of CGM data providers were generally aligned with the baseline characteristics of the overall population (Supplementary Table 2).

Table 1

Participant demographics and background characteristics

CGM subset (N = 926)*
Age (years) 41.9 ± 15.7 
Age ≥65 years 110 (11.9) 
Female 674 (72.8) 
White 886 (95.7) 
BMI (kg/m2) 27.5 ± 6.3 
Duration since diagnosis of type 1 diabetes (years) 25.0 ± 15.5 
Insulin delivery method  
 MDI 162 (17.5) 
 Sensor-augmented pump 293 (31.6) 
 AID 470 (50.8) 
Duration of CGM use  
 3 months to <1 year 55 (5.9) 
 1 to <3 years 245 (26.5) 
 3 to <5 years 245 (26.5 
 ≥5 years 381 (41.1) 
Duration of available CGM data in the past year (weeks)§ 41.6 ± 14.9 
 Minimum, maximum 1.6, 52.1 
Self-reported CGM wear frequency  
 <50% of time 0 (0) 
 50% of time 6 (0.6) 
 >50% to <100% of time 52 (5.6) 
 100% of time 867 (93.6) 
CGM subset (N = 926)*
Age (years) 41.9 ± 15.7 
Age ≥65 years 110 (11.9) 
Female 674 (72.8) 
White 886 (95.7) 
BMI (kg/m2) 27.5 ± 6.3 
Duration since diagnosis of type 1 diabetes (years) 25.0 ± 15.5 
Insulin delivery method  
 MDI 162 (17.5) 
 Sensor-augmented pump 293 (31.6) 
 AID 470 (50.8) 
Duration of CGM use  
 3 months to <1 year 55 (5.9) 
 1 to <3 years 245 (26.5) 
 3 to <5 years 245 (26.5 
 ≥5 years 381 (41.1) 
Duration of available CGM data in the past year (weeks)§ 41.6 ± 14.9 
 Minimum, maximum 1.6, 52.1 
Self-reported CGM wear frequency  
 <50% of time 0 (0) 
 50% of time 6 (0.6) 
 >50% to <100% of time 52 (5.6) 
 100% of time 867 (93.6) 

Data are means ± SD or n (%) unless otherwise indicated.

*Percentages across categories may not add to 100% due to rounding.

†BMI missing for 5 of 926 participants included in the analysis.

‡AID information missing for one pump user.

§CGM data sufficiency criterion: at least one segment of 14 consecutive days during which CGM was in use for ≥70% of the time among the 365 days prior to the date of data upload.

¶Data for frequency of CGM wear were missing for one participant.

Mean self-reported HbA1c was 6.6%, with 69.0% of participants reporting HbA1c <7% (Fig. 1A). Mean CGM-derived GMI was 6.9%, with 58.9% achieving GMI <7% (Fig. 1B). A strong correlation was observed between self-reported HbA1c and GMI (ρ [Spearman correlation coefficient] = 0.79). Both HbA1c <7% and GMI <7% were achieved by a numerically lower proportion of those who used MDI versus pump therapy and AID (Fig. 1A and B).

Figure 1

Participants achieving glycemic targets, measured according to HbA1c <7% (A) and GMI <7% (B), in the CGM subset (N = 926). aAID information missing for one pump user. bMost recent self-reported HbA1c value within the past 12 months; only participants who reported HbA1c within the past 12 months were included. cError bars reflect 95% CI.

Figure 1

Participants achieving glycemic targets, measured according to HbA1c <7% (A) and GMI <7% (B), in the CGM subset (N = 926). aAID information missing for one pump user. bMost recent self-reported HbA1c value within the past 12 months; only participants who reported HbA1c within the past 12 months were included. cError bars reflect 95% CI.

Close modal

TIR was derived for each participant with use of historical CGM data. Mean percent TIR was 66.6% for CGM+MDI, 68.0% for CGM+pump, and 76.1% for CGM+AID. Mean percent TITR was 44.3% for CGM+MDI, 46.4% for CGM+pump, and 52.2% for CGM+AID (Fig. 2A). The proportions who met target TIR of >70% were 50.0% for CGM+MDI, 49.5% for CGM+pump, and 72.1% for CGM+AID (Fig. 2B). Numerically higher proportions achieved TAR, TIR, and TBR targets with CGM+AID versus CGM with MDI or pump (Fig. 2B). To better understand a clinically relevant goal of meeting target TIR while avoiding low glucose values (i.e., meeting TBR [<70 mg/dL] goals), we evaluated a composite of these metrics; 25.9% of the CGM+MDI group, 27.0% CGM+pump, and 59.1% CGM+AID met this target (Fig. 2B). The majority (68.6%) met target glycemic variability (CV ≤36%), with 74.9% of the CGM+AID group achieving it (Supplementary Fig. 2).

Figure 2

Mean percent time spent with glucose in various ranges (A) and proportion of participants achieving various glucose time in range targets (B) by insulin delivery method in the CGM subset (N = 926). aAID information missing for one pump user. bError bars reflect 95% CI. TAR, >180 mg/dL; TIR, 70–180 mg/dL; TITR, 70–140 mg/dL; TBR, <70 mg/dL.

Figure 2

Mean percent time spent with glucose in various ranges (A) and proportion of participants achieving various glucose time in range targets (B) by insulin delivery method in the CGM subset (N = 926). aAID information missing for one pump user. bError bars reflect 95% CI. TAR, >180 mg/dL; TIR, 70–180 mg/dL; TITR, 70–140 mg/dL; TBR, <70 mg/dL.

Close modal

IAH was reported by 31.5% of participants (Supplementary Fig. 3). Similar rates of self-reported SHEs (one or more events, 16.4%; two or more, 9.8%) and CGM-derived SDPH events (one or more events, 22.8%; two or more, 12.7%) (Supplementary Fig. 4) were observed over the 12-month period. In assessing SHEs by insulin delivery method we found that the proportion of participants with one or more SHEs was 19.1%, 16.4%, and 15.5% in the CGM+MDI, CGM+pump, and CGM+AID groups, respectively (Fig. 3A), and the proportion of participants with one or more SDPH events was 26.5%, 31.7%, and 16.0% (Fig. 3A).

Figure 3

Proportion of participants with SHEs and proportion of participants with SDPH events in the CGM subset (N = 926) (A) and concordance between SHEs and SDPH events over 12 months (N = 924 [data on SHEs missing for two participants]) (B). Data in B are n (%). aSevere low blood glucose events: patients were unable to treat the events themselves and needed help from others. bAID information missing for one pump user. cAnnualized rate of SDPH events derived from the available CGM data. dNumber of self-reported SHEs over the past 12 months.

Figure 3

Proportion of participants with SHEs and proportion of participants with SDPH events in the CGM subset (N = 926) (A) and concordance between SHEs and SDPH events over 12 months (N = 924 [data on SHEs missing for two participants]) (B). Data in B are n (%). aSevere low blood glucose events: patients were unable to treat the events themselves and needed help from others. bAID information missing for one pump user. cAnnualized rate of SDPH events derived from the available CGM data. dNumber of self-reported SHEs over the past 12 months.

Close modal

Rates of self-reported SHE and SDPH events were concordant in 70% of participants (i.e., no SHEs/no SDPH events or one or more SHEs/one or more SDPH events) (Fig. 3B). Concordance was 82% for participants without IAH vs. 61% for participants with IAH (proportion with IAH: 31.5% [Supplementary Table 3]). A greater proportion of participants with no SHEs/no SDPH events was observed among those without IAH (71%) versus those with IAH (53%) (Supplementary Table 3). Participants reporting more SHEs were less likely to meet TBR and CV targets (e.g., 78.4%, 68.9%, 63.5%, and 67.9% achieved <4% TBR with no SHEs, one SHE, two to four SHEs, and five or more SHEs, respectively—with a similar pattern for proportion meeting the CV target) (Supplementary Table 4).

The results of this observational, retrospective study demonstrate that despite use of diabetes management technologies, many participants did not achieve glycemic targets and experienced SHEs as measured according to self-report and CGM metrics. CGM is the standard of care, and AID systems can help improve glycemia and reduce hypoglycemia, and some have posited that these devices have “solved” issues surrounding type 1 diabetes management. This simplistic assessment fails to recognize limitations of technology. While advanced technologies have permitted more individuals to reach targets, many patients continue to struggle to maintain glucose in ideal ranges, as highlighted in our data.

Although a greater proportion of participants using CGM+AID met HbA1c and CGM targets versus other insulin delivery methods, ∼28% using CGM+AID were unable to meet the TIR target of >70% of time spent with glucose in range 70–180 mg/dL. While improvements in diabetes-specific quality of life measures and overall treatment satisfaction have been observed with CGM and AID, the many requirements for handling these technologies (including device management, lifestyle adjustments, supply maintenance, meal announcement, attention to alarms, responsiveness to prompts, and data review)—may contribute to individuals not reaching glycemic targets (6–9). Moreover, there remain gaps between physiologic glucose homeostasis and current clinical glucose management options, such as exogenous insulin therapy, which is administered subcutaneously, regardless of method, and has a variable pharmacological profile.

In our study we examined concordance between self-reported and CGM-derived measures of glycemic control and hypoglycemia. Study results showed a higher proportion of people self-reporting achievement of HbA1c <7% than those achieving GMI <7%, potentially reflective of numerous factors that can contribute to this discrepancy (e.g., anemia, kidney disease, or genetic factors). Additionally, the data highlight the contrast between self-report of clinical testing for HbA1c and a value imputed from interstitial CGM readings for GMI. Despite these potential variables, there was a strong correlation between HbA1c and GMI, on a population level, and consistent with prior studies (10). Concordance of 70% was observed between self-reported and CGM-derived hypoglycemia measures. However, it is important to capture patient-report and CGM metrics, as they provide complementary information. Specifically, 18% of participants had SDPH captured by CGM and had no self-reported SHEs. This finding is consistent with a separate study demonstrating that the majority of sensor-detected hypoglycemic events were not reported (11). While this highlights the benefits and sensitivity of CGM technology, a proportion of SDPH events could also be “false hypoglycemia” due to biomechanical sensor compression, inaccuracy between sensor readings and blood glucose, or other device malfunction. Conversely, 12% of study participants reported at least one SHE and had no demonstrated SDPH; among those without IAH, a smaller proportion (9%) of participants experienced SHEs without SDPH. This interesting observation highlights the importance of clinical context in interpreting patient-reported as well as sensor-detected hypoglycemia.

Study limitations include a potential for sampling bias given that participants were enrolled from the U.S.-based T1D Exchange Registry, with a predominantly White, female, and privately insured cohort with a relatively high degree of technology adoption and access to specialist care. As such, study results may not fully represent outcomes in the general population with type 1 diabetes, and further investigation are needed of real-world conditions, which may present with even more stark findings. In addition, almost all participants (97%) reported using Dexcom Clarity for upload of CGM data; hence, the applicability of these data to other CGM brands is unknown. Nevertheless, these findings of persistent glycemic burden within an engaged and well-resourced cohort with largely at-goal glycemic metrics are particularly notable and indicate an ongoing need in characterizing these “at risk” subpopulations. Self-reported survey results may be subject to recall bias; however, the strong correlation between self-reported HbA1c and CGM-derived GMI values is reassuring regarding comparing self-reported measures with CGM-derived metrics. These data are for measurement of different constructs, and recall periods may not be equivalent. However, analyses comparing HbA1c in the prior 3 months and GMI calculations based on proximal 2-week data yielded similar results (12).

This study demonstrates that many participants using diabetes technologies remain unable to achieve clinical targets, have IAH, and experience significant hypoglycemia, according to self-report and CGM data. These findings highlight the need for both the continued improvement of AID systems and the evaluation of novel therapies to improve type 1 diabetes care and outcomes.

This article contains supplementary material online at https://doi.org/10.2337/figshare.27637839.

L.T. is currently affiliated with GlaxoSmithKline.

Acknowledgments. J.P. is an editor of Diabetes Care but was not involved in any of the decisions regarding review of the manuscript or its acceptance.

Funding and Duality of Interest. This study was funded by Vertex Pharmaceuticals. Medical writing support and editing support were provided under the guidance of the authors by Complete HealthVizion, funded by Vertex Pharmaceuticals. L.M.L. has received consulting fees from Boehringer Ingelheim, Medtronic, Tandem Diabetes Care, Dexcom, Arbor Biotechnologies, Janssen, Sanofi, Sequel, MannKind, and Vertex Pharmaceuticals. J.L.S. has received grants or contracts from Breakthrough T1D, the National Institute of Diabetes and Digestive and Kidney Diseases, Jaeb Center for Health Research, Insulet, Medtronic, Provention Bio, and Abbott paid to her institution; has received consulting fees from Abbott; has received payment or honoraria from Insulet, Medtronic Diabetes, and Zealand Pharma; has participated in advisory boards for Bigfoot Biomedical, Cecelia Health, Insulet, Medtronic Diabetes, Novo Nordisk, and Vertex Pharmaceuticals; and reports owning stock/stock options of StartUp Health T1D Moonshot. J.L., W.A.W., J.B., K.S.C., and D.F. were employees of T1D Exchange during the course of the study. L.T. is a board member for COJECO, was an employee at Vertex Pharmaceuticals during the course of the study, and is currently an employee at GlaxoSmithKline. J.G. has received grants or contracts from Avotres, Dompé farmaceutici, and Imcyse paid to his institution; has received consulting fees from Avotres, Imcyse, Dompé farmaceutici, Regeneron Pharmaceuticals, Vertex Pharmaceuticals, Diamyd Medical, and Current Health; has received payment or honoraria from Foundation for Advanced Education in the Sciences at the National Institutes of Health; and reports owning stock/stock options of Vertex Pharmaceuticals. T.L., K.H., and K.C. are employees at Vertex Pharmaceuticals and own stocks/stock options of Vertex Pharmaceuticals. J.P. has received consulting fees from Novo Nordisk, Sanofi, Eli Lilly, Carmot Therapeutics, Diasome, Kriya Therapeutics, and Biomea Fusion. R.B. has received consulting fees from Abbott Diabetes Care, Ascensia Diabetes Care, Bigfoot Biomedical, Dexcom, MannKind, Medtronic, Novo Nordisk, Sanofi, and United Healthcare paid to HealthPartners Institute; has received payment or honoraria from Sanofi and Vertex Pharmaceuticals paid to HealthPartners Institute; has received support to attend meetings from Abbott Diabetes Care, Ascensia Diabetes Care, CeQur, Eli Lilly, Embecta, MannKind, Novo Nordisk, Roche GmbH, Sanofi, Vertex Pharmaceuticals, and Zealand Pharma; and has participated in advisory boards for Abbott Diabetes Care, CeQur, Eli Lilly, Embecta, Hygieia, Roche GmbH, and Zealand Pharma. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. All authors contributed to the development of the study design. L.M.L., J.L., W.A.W., J.B., K.S.C., and D.F. acquired study data. L.M.L., J.L., W.A.W., J.B., K.S.C., D.F., L.T., T.L., and K.H. conducted the analyses of the study data. All authors were involved in interpreting the data, reviewing and drafting the manuscript, and approving the final version of the manuscript to be published. All authors agreed to be accountable for the work in the manuscript. K.H. and R.B. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in abstract form at the 59th Annual Meeting of the European Association for the Study of Diabetes, Hamburg, Germany, 2–6 October 2023.

Handling Editors. The journal editor responsible for overseeing the review of the manuscript was Mark A. Atkinson.

1.
Sherr
JL
,
Laffel
LM
,
Liu
J
, et al
.
Severe hypoglycemia and impaired awareness of hypoglycemia persist in people with type 1 diabetes despite use of diabetes technology: results from a cross-sectional survey
.
Diabetes Care
2024
;
47
:
941
947
2.
Rodbard
D
.
Continuous glucose monitoring: a review of recent studies demonstrating improved glycemic outcomes
.
Diabetes Technol Ther
2017
;
19
:
S25
S37
3.
Kravarusic
J
,
Aleppo
G
.
Diabetes technology use in adults with type 1 and type 2 diabetes
.
Endocrinol Metab Clin North Am
2020
;
49
:
37
55
4.
Gold
AE
,
MacLeod
KM
,
Frier
BM
.
Frequency of severe hypoglycemia in patients with type I diabetes with impaired awareness of hypoglycemia
.
Diabetes Care
1994
;
17
:
697
703
5.
Danne
T
,
Nimri
R
,
Battelino
T
, et al
.
International consensus on use of continuous glucose monitoring
.
Diabetes Care
2017
;
40
:
1631
1640
6.
Polonsky
WH
,
Hessler
D
,
Ruedy
KJ
,
Beck
RW
;
DIAMOND Study Group
.
The impact of continuous glucose monitoring on markers of quality of life in adults with type 1 diabetes: further findings from the DIAMOND randomized clinical trial
.
Diabetes Care
2017
;
40
:
736
741
7.
Lind
M
,
Polonsky
W
,
Hirsch
IB
, et al
.
Continuous glucose monitoring vs conventional therapy for glycemic control in adults with type 1 diabetes treated with multiple daily insulin injections: the GOLD randomized clinical trial
.
JAMA
2017
;
317
:
379
387
8.
Crabtree
TSJ
,
Griffin
TP
,
Yap
YW
, et al.;
ABCD Closed-Loop Audit Contributors
.
Hybrid closed-loop therapy in adults with type 1 diabetes and above-target HbA1c: a real-world observational study
.
Diabetes Care
2023
;
46
:
1831
1838
9.
Nørgaard
K
,
Ranjan
AG
,
Laugesen
C
, et al
.
Glucose monitoring metrics in individuals with type 1 diabetes using different treatment modalities: a real-world observational study
.
Diabetes Care
2023
;
46
:
1958
1964
10.
Sterner Isaksson
S
,
Imberg
H
,
Hirsch
IB
, et al
.
Discordance between mean glucose and time in range in relation to HbA1c in individuals with type 1 diabetes: results from the GOLD and SILVER trials
.
Diabetologia
2024
;
67
:
1517
1526
11.
Divilly
P
,
Martine-Edith
G
,
Mahmoudi
Z
, et al.;
Hypo-RESOLVE Consortium
.
Majority of sensor hypoglycemia is not detected by people living with diabetes—Hypo-METRICS Study
. Diabetes 2023;72(Suppl. 1):250
12.
Riddlesworth
TD
,
Beck
RW
,
Gal
RL
, et al
.
Optimal sampling duration for continuous glucose monitoring to determine long-term glycemic control
.
Diabetes Technol Ther
2018
;
20
:
314
316
Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at https://www.diabetesjournals.org/journals/pages/license.