Pediatric Medicaid Patients With Type 1 Diabetes Benefit From Continuous Glucose Monitor Technology

Abstract

Background:

We determined the uptake rate of continuous glucose monitors (CGMs) and examined associations of clinical and demographic characteristics with CGM use among patients with type 1 diabetes covered by Colorado Medicaid during the first two years of CGM coverage with no out-of-pocket cost.

Method:

We retrospectively reviewed data from 892 patients with type 1 diabetes insured by Colorado Medicaid (Colorado Health Program [CHP] and CHP+, Colorado Medicaid expansion). Demographics, insulin pump usage, CGM usage, and hemoglobin A1c (A1c) were extracted from the medical record. Data downloaded into CGM software at clinic appointments were reviewed to determine 30-day use prior to appointments. Subjects with some exposure to CGM were compared to subjects never exposed to CGM, and we examined the effect of CGM use on glycemic control.

Results:

Twenty percent of subjects had some exposure to CGM with a median of 22 [interquartile range 8, 29] days wear. Sixty one percent of CGM users had >85% sensor wear. Subjects using CGM were more likely to be younger (P < .001), have shorter diabetes duration (P < .001), and be non-Hispanic White (P < .001) than nonusers. After adjusting for age and diabetes duration, combined pump and CGM users had a lower A1c than those using neither technology (P = .006). Lower A1c was associated with greater CGM use (P = .002) and increased percent time in range (P < .001).

Conclusion:

Pediatric Medicaid patients successfully utilized CGM. Expansion of Medicaid coverage for CGM may help improve glycemic control and lessen disparities in clinical outcomes within this population.

Keywords

continuous blood glucose monitor low income Medicaid pediatric type 1 diabetes technology and diabetes

Introduction

Type 1 diabetes is a common chronic condition, affecting 1 in 300 children.¹ Intensive glycemic control in type 1 diabetes decreases the risk of long-term microvascular complications, and frequent blood glucose (BG) monitoring is an important factor in attaining better glucose control.² Continuous glucose monitors (CGMs) have been commercially available for over a decade to augment the treatment of type 1 diabetes. Consistent CGM use greater than six days per week is associated with improved hemoglobin A1c (A1c) in the pediatric population,^3,4 and studies have consistently shown better A1c trends among children who use CGM.^5-8 Additional benefits include decreased glucose variability, risk of hypoglycemic events, and diabetic ketoacidosis.^3,9,10 The benefits of CGM will likely expand as more insulin pump systems automate insulin delivery based on CGM readings.^11,12

Major barriers to starting CGM are cost and inconsistent insurance coverage.^4,10 Low income and minority patients have higher A1cs than their counterparts,^13,14 and CGM coverage by Medicaid for low income children varies by state. Therefore, the clinical benefits of CGM are disproportionally more available for Caucasian and privately insured patients. Although children with lower socioeconomic status were less likely to continue CGM use six months after implementation with older CGM models, more recent data show 70% of low income Medicaid children had continued use at six months with newer systems.^10,15-17 In Colorado, the income requirement for Medicaid eligibility (Colorado Health Program [CHP] and CHP+, the Colorado Medicaid expansion program) is household income less than 260% of the federal poverty level. For these patients, CGM first became a covered benefit in July 2015¹⁸ and there were no requirements aside from provider prescription to obtain a CGM for children with type 1 diabetes.¹⁹

The purpose of this study was to determine the CGM uptake rate and examine associations of demographic and clinical characteristics with CGM use among patients with type 1 diabetes covered by Colorado Medicaid during first two years of CGM coverage.

Methods

This study was a retrospective chart review conducted at a large tertiary care academic practice in Aurora, Colorado. Participants with CHP or CHP+ insurance were included starting at their first outpatient encounter between July 1, 2015 and July 31, 2017. Data were obtained for all subsequent encounters until the end of the study period or until the first encounter when their insurance was no longer CHP or CHP+. Data for patients who subsequently returned to public insurance following insurance change were not collected. The A1c was measured on the DCA2000+ (Siemens/Bayer, Munich, Germany) as part of routine clinical care and extracted from the medical record for each encounter. Age, diabetes duration, sex, race/ethnicity, pump usage (yes/no), and CGM possession (yes/no) were also extracted from the medical record. There were no clinic policies (eg, number of BG checks per day) for CGM prescription, and providers relied on assessing patient readiness. Colorado Medicaid, however, did require documentation of hypoglycemia, either by meter download or oral report, to authorize the prescription. It is routine practice in the clinic to download 14 days of BG meter, insulin pump, and CGM data at each outpatient encounter and to record percent of BG readings below, within, and above the target range. Target range was defined as 70 to 180 mg/dL [3.9-10 mmol/L].⁹ Continuous glucose monitor data stored in either Medtronic Carelink (Los Angeles, CA, United States), Dexcom Studio (San Diego, CA, United States), or Dexcom Clarity (San Diego, CA, United States) were reviewed to record the number of days CGM was used in 30 days prior to the encounter date. This study was approved by the Colorado Multiple Institutional Review Board.

Data Analysis

Variables were assessed for normality using the Kolmogorov-Smirnov test. Continuous variables were compared using t-tests or Kruskal-Wallis Rank Sum tests, and chi-squared or Fisher’s exact tests were used for categorical variables. We calculated the mean A1c for each subject and report the median of these average A1c values in the descriptive statistics. For the CGM nonexposed group, we use all time points in the study. For the CGM exposed group, we report A1c both before and after starting CGM. The effect of technology status (no technology, pump only, CGM only, and both pump and CGM) on A1c was examined using a linear mixed effects model with random intercept for subject adjusted for age and diabetes duration.

We also examined the CGM exposed group independently of the nonexposed group. We determined the median days of sensor wear in 30 days prior to each encounter and the percentage of patients with at least 85% sensor wear (25/30 days). This threshold for wear has been shown in the previous studies to correlate with improved glycemic control.^3,4 We determined the median sensor wear in CGM users at one year of CGM use (300-420 days from first encounter with CGM). We determined the percent of time CGM users spent below, within, and above target range as per consensus guidelines.⁹ A mixed model analysis was performed to determine factors that would predict successful use of CGM defined as >85% wear in 30 days prior to clinic visits. The effect of percent time in range and percent CGM use on A1c was examined using linear mixed effect models with a random intercept for subject. When modeling A1c with percent CGM use as a predictor, we adjusted for age, diabetes duration, pump use, and race/ethnicity.

Analyses were performed using R version 3.6.0 (Vienna, Austria) and descriptive statistics were compared using the “tableone” package.²⁰

Results

In the two-year period studied, 892 patients met the inclusion criteria. Of these, 19.8% had some exposure to CGM. The descriptive characteristics of never CGM users and CGM users are shown in Table 1. The median A1c prior to starting CGM was 8.5% [interquartile range (IQR) 7.9, 9.7] (69 mmol/mol [63-83]). Participants with CGM exposure were more likely to be non-Hispanic White (P < .001) and less likely to be Hispanic (P < .001) than nonusers. Subjects with some exposure to CGM were younger and had shorter diabetes duration (P < .001).

Table 1.

Description of the Population.

	Never used CGM (N = 715)	Some exposure to CGM (N = 177)	P-value
Age at first visit (y)	13.1 ± 4.26	10.7 ± 4.5	<.001
Diabetes duration (y)	8.1 [4.9, 10.9]	5.8 [3.0, 9.2]	<.001
Males	349 (48.8)	82 (46.3)	.61
Pump users	278 (38.9)	130 (73.4)	<.001
Race/ethnicity			<.001
Non-Hispanic White	292 (40.8)	111 (62.7)	<.001
Hispanic	264 (36.9)	40 (22.6)	<.001
Black	57 (8.0)	7 (4.0)	.07
American Indian	9 (1.3)	0 (0.0)	.22
Asian	4 (0.6)	3 (1.7)	.15
Mixed race	18 (2.5)	4 (2.3)	1
Other/unknown	70 (9.8)	12 (6.7)	.27
A1c before CGM (%)	9.3 [8.4, 10.7]	8.5 [7.9, 9.7]	<.001
(mmol/mol)	78 [68, 93]	69 [63, 83]
A1c after possession of CGM (%)	-	8.6 [7.9, 9.7]	<.001^a
(mmol/mol)	-	70 [63, 83]
Age <6 y (n = 61)	CGM % time below target range (median [IQR])	2.00 [1.0, 4.0]	-
	CGM % time in target range (mean (SD))	41.8 (15.4)	-
	CGM % time above target range (mean (SD))	54.75 (17.3)	-
Age 6-<13 y (n = 208)	CGM % time below target range (median [IQR])	2.00 [0.5, 3.0]	-
	CGM % time in target range (mean (SD))	39.4 (16.0)	-
	CGM % time above target range (mean (SD))	58.0 (17.1)	-
Age 13-17 y (n = 138)	CGM % time below target range (median [IQR])	1.00 [0.0, 3.0]	-
	CGM % time in target range (mean (SD))	35.20 (20.3)	-
	CGM % time above target range (mean (SD))	62.28 (21.3)	-
Age >18 y (n = 39)	CGM % time below target range (median [IQR])	2.00 [0.5, 5.0]	-
	CGM % time in target range (mean (SD))	26.56 (17.0)	-
	CGM % time above target range (mean (SD))	69.92 (18.6)	-

Abbreviations: CGM, continuous glucose monitor; IQR, interquartile range.

Statistics given as N (%), mean ± standard deviation, or median [interquartile range]. Target range defined as 70-180 mg/dL [3.9-10 mmol/L].

Compared to A1c of the never used continuous glucose monitor group.

Among CGM users, 27% (47/177) used CGM with multiple daily injections, and 73% (130/177) used CGM with an insulin pump. Median CGM use was 22 [8, 29] out of 30 days, and 61% of patients had at least 85% wear (25/30 days) in 30 days prior to clinic visits. Median percent time in hyperglycemia was 59.9% [46.0-75.4], time in target range was 37.5% [22.5, 49.1], and time in hypoglycemia was 2.0% [0.7, 3.8] (Table 1). The median sensor wear at one year after the first CGM encounter was 86.7% [48.3, 96.7]. The median sensor wear at one year after the first CGM encounter demonstrates that not all participants had 85% wear over the entire study period; however, mean TIR and wear percent were relatively stable across visits (Figure 1). Using logistic regression to predict successful use of CGM, age (P = .004) and race (P = .013) predicted CGM use >85%. On average, each year increase in age resulted in a 10.33% decrease in the odds of successful CGM use (P = .004). Hispanic participants were 66.5% less likely to successfully use CGM compared to non-Hispanic Whites (P = .003). Days since last visit were not associated with successful CGM use (P = .914).

Figure 1.

Percent time in target range (TIR) and continuous glucose monitor wear by visit number.

Overall, subjects who used CGM had a lower A1c prior to starting CGM compared to non-CGM users (P < .001). Subjects with some exposure to CGM also had a significantly lower A1c (median 8.6% [7.9, 9.7], 70 mmol/mol [63, 83]) after CGM exposure compared to those never exposed to CGM (9.3% [8.4, 10.7], 78 mmol/mol [68, 93]) (P < .001). The mixed model showed that when controlling for age and diabetes duration, A1c was significantly lower for patients who were exposed to CGM and were using an insulin pump (9.1% ± 0.1%, 76 ± 0.85 mmol/mol) compared to subjects utilizing neither a pump nor CGM (A1c 9.7% ± 0.1%, 83 ± 0.85 mmol/mol) (P = .006), but none of the other comparisons were significant (Table 2).

Table 2.

Mixed Model Comparing A1c in Subjects Never on Continuous Glucose Monitor to A1c in Subjects after Continuous Glucose Monitor Exposure by Technology Status, Adjusting for Age, and Diabetes Duration.

	No pump (no CGM) n = 437	Pump only (no CGM)n = 278	CGM only (no pump)n = 47	CGM and pumpn = 130
Least squares mean A1c
%	9.7 ± 0.1	9.5 ± 0.1	9.4 ± 0.2	9.1 ± 0.1
mmol/mol	83 ± 0.9	80 ± 0.9	79 ± 1.7	76 ± 0.9
P-value compared to CGM and pump	.006	.17	.11	-
P-value compared to CGM only (no pump)	.52	.99	-	-
P-value compared to pump only (no CGM)	.13	-	-	-

Abbreviation: CGM, continuous glucose monitor.

Statistics given as mean ± standard error. P-values are adjusted for multiple comparisons using Tukey’s Honestly Significant Difference test.

Among CGM users, lower A1c was correlated with greater CGM use and increased percent time in range. After adjusting for race, diabetes duration, and pump use, each ten-point increase in percent CGM use was associated with a decrease in A1c of 0.04% ± 0.01% (0.4 ± 0.1 mmol/mol) (P < .01). For each 1% increase in time in target range (70-180 mg/dL, 3.9-10 mmol/mol), A1c decreased by 0.04% ± 0.01% (0.34 ± 0.1 mmol/mol) (P < .001) (Figure 2(a) and (b)).

Figure 2.

Best fit lines for A1c modeled against percent continuous glucose monitor use (a) and percent time in range (b) controlled for age, diabetes duration, pump use, and race/ethnicity.

Discussion

In this retrospective chart review of children with Colorado Medicaid insurance, 19.8% had exposure to CGM within the first two years of coverage with high sensor wear when cost is eliminated as a barrier to implementation. Unlike privately insured patients who may have additional copays or who must meet certain other conditions to have insurance coverage for CGM technology, pediatric patients with type 1 diabetes covered under the Colorado Medicaid program experience no out-of-pocket cost for CGM use and only need a provider prescription. Elimination of the cost barrier in our cohort allowed for successful clinical use of CGM.

Factors associated with CGM use in our study were similar to those found in larger studies.^6,14 Between 2015 and 2017, CGM use in the T1D Exchange (adults and pediatrics) increased from 18% to 31%, while use in the pediatric population was higher.⁶ Continuous glucose monitor users in our study were significantly younger than those not using CGM. In this same time period, the T1D Exchange found CGM use decreased with age among pediatric patients with the highest use in those <6 years of age and the lowest use in those 13 to 18 years of age.⁶ We postulate that this is likely because caregiver remote monitoring available on certain CGM models is highly desired by families of younger children.

Also consistent with the previous data,^8,14,21 CGM users in our cohort were more likely to be non-Hispanic White. Moreover, non-Hispanic White CGM users were more likely than their Hispanic counterparts who used CGM to achieve >85% wear. The failure to detect these disparities between non-Hispanic White patients and those from other racial/ethnic groups was likely due to relatively small numbers of non-White patients in our cohort. It is unclear to what extent these disparities may be due to cultural factors or prescribing bias, but it is important to recognize these challenges to prescribing CGMs and supporting patients in their use.

Patients with some exposure to CGM over the study period had a significantly lower A1c compared to subjects not using CGM. This is similar to findings from the T1D Exchange and Diabetes-Patienten-Verlaufsdokumentation (DPV) registries comparison.⁸ When data were adjusted for age and diabetes duration, the comparison of the group utilizing both CGM and an insulin pump to the group using no technology remained significant. This should be interpreted with caution however, as there is a clear selection bias, especially among early adopters of CGM. Subjects who used CGM had a lower A1c prior to starting CGM compared to the non-CGM users. Therefore, there may be other factors such as better baseline adherence or younger age contributing to improved glycemic control in the CGM exposed group.^15-17 Notably, there was no change when comparing the A1c before CGM to the A1c after CGM. This is an unexpected finding from the study. A1c tends to increase between ages 7 and 17 years,⁶ and most subjects were within that age range during 2 years of the study. Therefore, it will be important to further explore the age-independent relationship between CGM use and A1c. As age increased, CGM use declined. This was an expected finding, as management habits and glycemic trends both deteriorate from childhood through adolescence as diabetes management responsibility gradually shifts from the parents to the child. Time interval between outpatient visits was not associated with CGM use percent, suggesting that patients who struggle to use the CGM successfully may need additional supports beyond in-person education. This is further suggested by the finding that CGM use and time in range stayed fairly consistent over time in the study. There may be an initial adjustment to CGM use that may decrease sustained use but those who have consistent wear initially are more likely to sustain use over time.

Lower A1c in the CGM exposed group was significantly correlated with longer sensor wear. Over 60% of users had >85% sensor wear in 30 days prior to their encounters, which corresponds to 6/7 days, a measure of sensor wear was chosen because it has been shown to be associated with improved glycemic control.^3,4 Sustained use of CGM was also high, demonstrating that patients experienced sustained benefits of CGM technology. This contrasts with the previous studies which suggest that children with lower socioeconomic status were less likely to sustain CGM use six months after implementation.^15-17

We found that every 10% increase in time in target range was associated with an A1c decrease of 0.4%, which is slightly less than what has been reported previously.^22,23 This difference may be that unlike these prior studies, our population was pediatric, had a relatively higher mean A1c, and was composed of a larger percent of patients with ethnic minority background. The relatively higher A1c may indicate that some of our patients spent much more time with blood sugar >250 mg/dL, undoing some of the effect of the time in target range to achieve target range A1c. The relationship between time in range and sensor wear stayed consistent over time.

Since the time of the study period, new CGM models have improved comfort, ease of use, accuracy, and some are approved for independent dosing decisions.²⁴ Moreover, CGM technology is now used in hybrid-closed loop systems to automate insulin delivery, thus providing greater benefit than was available during the study period. Also since the time of the study, our clinic has adopted a new process to ensure that all patients recently diagnosed with diabetes are encouraged to attend a CGM class with the goal of increasing early CGM uptake and uniformly offering these tools to all patients regardless of their demographic.

Strengths of this study include the large sample size from a large practice which can be used to estimate the pediatric type 1 diabetes population covered by Colorado Medicaid. Additionally, there is a definite time point when CGM became a covered benefit for Colorado Medicaid, and this population is unlikely to afford the out-of-pocket costs allowing us to examine use when cost is eliminated as a barrier. This allows for the examination of a population underrepresented in past CGM studies.⁹

Limitations of this study include the single center, retrospective design. The retrospective design limits the ability to obtain qualitative data regarding patient/parent barriers to starting CGM as well as the ability to obtain satisfaction data for CGM use. Additionally, because the end of the date range was July 2017, this analysis did not include patients using hybrid closed loop insulin pump systems. The demographic makeup of the Colorado Medicaid population studied may limit generalizability to other state Medicaid populations.

Conclusion

This study demonstrates that pediatric patients with type 1 diabetes who are covered by Colorado Medicaid at our center have similar benefits from CGM as previously demonstrated across the pediatric population.^6,8,14,21 However, Medicaid does not cover CGM in all states.¹⁹ Given these results, we support the expansion of CGM coverage by Medicaid in other states. Barriers to CGM use in the Medicaid population clearly extend beyond cost. Age and racial disparities still exist, with CGM users more likely to be younger and non-Hispanic White. Further studies are warranted to identify these barriers and how best to address them. With advances in CGM accuracy, wearability, and applications, the clinical benefits are rapidly expanding. Failure to expand CGM coverage to the low-income population covered by Medicaid willing to utilize this technology may lead to increased disparities in clinical outcomes.

Footnotes

Declaration of Conflicting Interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: SJR and GTA designed the study with assistance from GPF. SJR and AC performed the research. SJR wrote and revised the manuscript. TV and LP performed the statistical analysis and edited the manuscript. AC, GPF, and GTA revised and edited the manuscript. All authors approved the final manuscript. SJR and GPF have funding as listed above. The other authors have no conflicts of interest. There has been no previous publication of this article or its abstract.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: SJR received the financial support from National Institutes of Health National Institute of Diabetes and Digestive Kidney Diseases (T32 DK063687). GPF received the financial support from Research support from Medtronic, Dexcom, Abbott, Tandem, and Insulet; and consulting fees from Medtronic, Dexcom, Abbott, Tandem, and Insulet. The effort was supported by National Institutes of Health National Institute of Diabetes and Digestive Kidney Diseases K12 award (K12DK094712).

ORCID iD

Guy T. Alonso

References

SEARCH Study Group. SEARCH for diabetes in youth: a multicenter study of the prevalence, incidence and classification of diabetes mellitus in youth. Control Clin Trials. 2004;25(5):458-471.

Diabetes Control and Complications Trial Research Group, Nathan

Genuth

, et al. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med. 1993;329(14):977-986.

Juvenile Diabetes Research Foundation Continuous Glucose Monitoring Study Group. Effectiveness of continuous glucose monitoring in a clinical care environment: evidence from the Juvenile Diabetes Research Foundation continuous glucose monitoring (JDRF-CGM) trial. Diabetes Care. 2010;33(1):17-22.

Juvenile Diabetes Research Foundation Continuous Glucose Monitoring Study Group, Beck

Buckingham

, et al. Factors predictive of use and of benefit from continuous glucose monitoring in type 1 diabetes. Diabetes Care. 2009;32(11):1947-1953.

Mulinacci

Alonso

Snell-Bergeon

Shah

. Glycemic outcomes with early initiation of continuous glucose monitoring system in recently diagnosed patients with type 1 diabetes. Diabetes Technol Ther. 2019;21(1):6-10.

Foster

Beck

Miller

, et al. State of type 1 diabetes management and outcomes from the T1D exchange in 2016-2018. Diabetes Technol Ther. 2019;21(2):66-72.

Miller

Foster

Beck

, et al. Current state of type 1 diabetes treatment in the U.S.: updated data from the T1D Exchange clinic registry. Diabetes Care. 2015;38(6):971-978.

DeSalvo

Miller

Hermann

, et al. Continuous glucose monitoring and glycemic control among youth with type 1 diabetes: international comparison from the T1D exchange and DPV initiative. Pediatr Diabetes. 2018;19(7):1271-1275.

Danne

Nimri

Battelino

, et al. International consensus on use of continuous glucose monitoring. Diabetes Care. 2017;40(12):1631-1640.

10.

Prahalad

Addala

Buckingham

Wilson

Maahs

. Sustained continuous glucose monitor use in low-income youth with type 1 diabetes following insurance coverage supports expansion of continuous glucose monitor coverage for all. Diabetes Technol Ther. 2018;20(9):632-634.

11.

Bekiari

Kitsios

Thabit

, et al. Artificial pancreas treatment for outpatients with type 1 diabetes: systematic review and meta-analysis. BMJ. 2018;361:k1310.

12.

Weisman

Bai

Cardinez

Kramer

Perkins

. Effect of artificial pancreas systems on glycaemic control in patients with type 1 diabetes: a systematic review and meta-analysis of outpatient randomised controlled trials. Lancet Diabetes Endocrinol. 2017;5(7):501-512.

13.

Delamater

Shaw

Applegate

, et al. Risk for metabolic control problems in minority youth with diabetes. Diabetes Care. 1999;22(5):700-705.

14.

Wong

Foster

Maahs

, et al. Real-time continuous glucose monitoring among participants in the T1D exchange clinic registry. Diabetes Care. 2014;37(10):2702-2709.

15.

Juvenile Diabetes Research Foundation Continuous Glucose Monitoring Study Group, Tamborlane

Beck

, et al. Continuous glucose monitoring and intensive treatment of type 1 diabetes. N Engl J Med. 2008;359(14):1464-1476.

16.

Chase

Beck

Xing

, et al. Continuous glucose monitoring in youth with type 1 diabetes: 12-month follow-up of the Juvenile Diabetes Research Foundation continuous glucose monitoring randomized trial. Diabetes Technol Ther. 2010;12(7):507-515.

17.

Giani

Snelgrove

Volkening

Laffel

. Continuous glucose monitoring (CGM) adherence in youth with type 1 diabetes: associations with biomedical and psychosocial variables. J Diabetes Sci Technol. 2017;11(3):476-483.

18.

Colorado

. https://www.colorado.gov/pacific/hcpf/colorado-medicaid. Accessed June 2019.

19.

Financing CDoHCPa. Early and periodic screening, diagnostic and treatment. https://www.colorado.gov/pacific/hcpf/early-and-periodic-screening-diagnostic-and-treatment-epsdt. Accessed June 2019.

20.

Yoshida

. tableone: Create ‘Table 1’to Describe Baseline Characteristics. R package version 0.10.0 2019. https://CRAN.R-project.org/package=tableone. Accessed June 2019.

21.

Lado

Lipman

. Racial and ethnic disparities in the incidence, treatment, and outcomes of youth with type 1 diabetes. Endocrinol Metab Clin North Am. 2016;45(2):453-461.

22.

Beck

Bergenstal

Cheng

, et al. The relationships between time in range, hyperglycemia metrics, and HbA1c. J Diabetes Sci Technol. 2019;13(4):614-626.

23.

Vigersky

McMahon

. The relationship of hemoglobin A1C to time-in-range in patients with diabetes. Diabetes Technol Ther. 2019;21(2):81-85.

24.

Forlenza

Argento

Laffel

. Practical considerations on the use of continuous glucose monitoring in pediatrics and older adults and nonadjunctive use. Diabetes Technol Ther. 2017;19(S3):S13-S20.