Sage Journals: Discover world-class research

Abstract

Background:

Real-time continuous glucose monitoring (rtCGM) improves glycemic control in type 1 diabetes (T1D) patients treated with continuous subcutaneous insulin infusion (CSII). However, the benefits of rtCGM in T1D patients treated with multiple daily insulin injection (MDI) therapy has not been well studied. We explored the effects of rtCGM versus self-monitoring of blood glucose (SMBG) on clinical outcomes within a large T1D population treated with either CSII or MDI therapy.

Methods:

This retrospective, longitudinal analysis utilized datasets from T1D patients enrolled in a commercial health plan to assess changes in HbA1c in 187 naïve to rtCGM users and 6260 SMBG users. Propensity score modeling was used to assess inpatient admissions, emergency room (ER) visits in 1130 patients (565 rtCGM, 565 SMBG). Differences in HbA1c reduction (rtCGM+MDI vs rtCGM+CSII) were evaluated.

Results:

Larger, clinically meaningful HbA1c reductions were seen among rtCGM versus SMBG users: –0.5% (P = .004) versus −0.2% (P < .0001); 0.3% diff in diff, P = .03. All-cause inpatient admissions were lower for rtCGM users: –42%, P = .013. Emergency room visits coded for diabetic ketoacidosis (DKA) were four times higher for SMBG patients than rtCGM patients: 17 versus 4, P = .0318. HbA1c reductions were most notable with rtCGM+MDI versus rtCGM+CSII treatment: –0.6% (P = .01) versus −0.3% (P = .16).

Conclusions:

Use of rtCGM in T1D patients facilitates greater HbA1c improvements and reduced health care system utilization compared with traditional SMBG use regardless of insulin administration method. Treatment with rtCGM in conjunction with MDI confers similar or greater glycemic benefits without the additional costs associated with CSII therapy.

Keywords

type 1 diabetes rtCGM continuous glucose monitoring SMBG CSII insulin pump

With recent advances in accuracy, reliability and usability, real-time continuous glucose monitoring (rtCGM) is now recognized as a critical component of diabetes self-management among many individuals with insulin-treated diabetes.¹ Unlike traditional blood glucose monitoring systems, which display only “point-in-time” glucose values, rtCGM devices provide numerical and graphical information about the current glucose level, glucose trends and the direction/velocity of changing glucose. Devices with programmable alerts/alarms that warn users of current and/or impending high or low glucose offer additional safety advantages.^2-5

Numerous studies have shown that use of rtCGM improves glycemic control in children and adults with type 1 (T1D).^3,6-15 Importantly, the benefits of rtCGM are seen primarily in individuals who wear their devices frequently and act appropriately on the glucose data provided.^3,6,16-18 Most of the evidence supporting rtCGM use comes from studies conducted primarily in patients using continuous subcutaneous insulin infusion (CSII);^7,14,18-21 however, results from a recent study of 20 adult T1D patients suggest that rtCGM may confer the same glycemic benefits to patients treated with multiple daily insulin injection (MDI) therapy.²²

The large financial burden of diabetes is growing and unsustainable, and clinicians are challenged to select treatment regimens that are both efficacious and cost-effective. Although new evidence suggests that rtCGM use in combination with CSII is more cost effective than CSII, alone,^23,24 the cost-effectiveness of rtCGM use with MDI therapy is largely unknown. However, given its enhanced decision support and the resulting improvements in glycemic control, rtCGM may be even more cost-effective than CSII because it eliminates the initial and ongoing costs of insulin pump use.

To explore this hypothesis, we analyzed the effects of rtCGM compared with self-monitoring of blood glucose (SMBG) use on clinical outcomes and utilization of health care services within a large population of T1D patients treated with either CSII or MDI therapy.

Methods

Design

This retrospective database analysis assessed the impact of rtCGM (G4 Platinum, Dexcom, Inc, San Diego, CA, USA) compared with traditional self-monitoring of blood glucose (SMBG) in a large cohort of T1D patients. The primary outcome measure was change in HbA1c between and within study groups, according to insulin delivery method (CSII vs MDI). Secondary outcomes included within- and between-group differences in hospitalizations and emergency room (ER) visits.

Data for the study were obtained from the Optum Research Database (Optum, Eden Prairie, MN, USA), which contains eligibility, pharmacy claims, medical claims and laboratory data for more than 14 million enrollees in fully insured and self-funded health care plans. Medical and demographic information, including diagnosis, utilization of health care services (eg, inpatient admissions, emergency room visits, pharmacy costs), age, gender and geographic regions were obtained from health plans’ administrative records for this study.

The identification period for eligible patients was from November 2012 through December 2013. The index date for each patient was the date of the first claim for initiation of either rtCGM or SMBG at a frequency of ≥4 test strips per day. The baseline period for each group was one year prior the index date; whereas, the measurement period was a year following the index date, including the index date itself.

Analysis

The study population included patients with a diagnosis code for T1D, continuous enrollment in the health plan and at least one claim for insulin during the study period. Patients who were pregnant or had prior experience with rtCGM were excluded from all analyses.

The primary analysis included patients who had at least one HbA1c value documented in both the baseline and the measurement periods. Study patients were divided into two groups: patients who initiated rtCGM use (rtCGM group) with the Dexcom G4 Platinum system (Dexcom, Inc, San Diego, CA, USA) and patients documented use of self-monitoring of blood glucose (SMBG group) at a frequency of ≥4 test strips per day within the baseline period as indicated by medical claims. Daily test strip use was calculated as the quantity of test strips dispensed divided by the days until the next prescription fill date. To control for the baseline difference of the two groups, the analysis was also stratified by baseline HbA1c value; thus, baseline matching or continuous enrollment was not required. Changes in HbA1c levels between the rtCGM and SMBG groups, overall and by insulin delivery method (CSII vs MDI), were compared; t-tests were performed to determine the statistical significance of between-group differences.

For the secondary analysis, propensity score matched analysis was performed to reduce selection bias due to imbalances in study covariates. The two groups were extensively matched on baseline per-member per-month (PMPM) medical and pharmacy costs, gender, region, Charlson Index Score and sixteen comorbidity Charlson indices.²⁵ Patients in the rtCGM group were matched to those in the SMBG group in a 1:1 ratio, based on the resultant propensity score probabilities. However, patients were not matched for HbA1c due to the relatively smaller number rtCGM patients with baseline and measurement period values (Table 1). The number of coded incidents of diabetic ketoacidosis (DKA) and hypoglycemia were calculated.

Table 1.

Demographic Characteristics.

	Prematch statistics				Postmatch statistics
	SMBG	CGM	Diff	P Value	SMBG	CGM	Diff	P Value
N	22 354	571			565	565
PMPM medical costs	$803	$879	$76	.10	$839	$869	$30	.67
PMPM pharmacy costs	$524	$569	$45	<.01	$545	$562	$17	.41
% female	45.4%	46.4%	1.0%	.64	44.1%	46.5%	2.5%	.40
Age (mean)	44.99	36.07	−8.92	<.01	38.02	36.22	−1.80	.08
≤ 20	15.3%	22.9%	7.7%	<.01	21.2%	22.5%	1.2%	.61
21 to 30	8.8%	14.2%	5.4%	<.01	13.8%	14.3%	0.5%	.80
31 to 40	10.4%	20.0%	9.5%	<.01	20.0%	20.0%	0.0%	1.00
41 to 50	17.8%	18.6%	0.7%	.65	19.8%	18.6%	−1.2%	.60
51 to 60	27.1%	17.0%	−10.1%	<.01	16.3%	17.2%	0.9%	.69
61 to 64	11.4%	4.2%	−7.2%	<.01	3.7%	4.2%	0.5%	.65
≥ 65	9.2%	3.2%	−6.0%	<.01	5.1%	3.2%	−1.9%	.10
Overall Charlson Index Score	1.89	1.67	−0.22	<.01	1.71	1.66	−0.05	.52
Diabetes complications	24.2%	25.6%	1.4%	.44	23.4%	25.5%	2.1%	.41
Congestive heart failure	5.0%	1.8%	−3.3%	<.01	1.9%	1.8%	−0.2%	.83
Cerebrovascular disease	4.8%	2.6%	−2.2%	.00	2.3%	2.7%	0.4%	.70
Chronic pulmonary disease	11.7%	10.5%	−1.2%	.36	11.0%	9.7%	−1.2%	.49
Peripheral vascular disease	4.5%	0.9%	−3.6%	<.01	1.6%	0.9%	−0.7%	.28

Summary statistics for continuous variables are summarized using mean and SD values. Analyses were conducted using SAS version 9.2 (SAS Institute Inc, Cary, NC, USA).

Results

Primary Analysis

A total of 6467 patients (187 rtCGM; 6280 SMBG) were included in the primary analysis. The distribution of baseline values in the two groups was similar (Figure 1).

Figure 1.

Percentage distribution of baseline HbA1c values.

The HbA1c reductions seen in the rtCGM group were significantly greater (P = .0305) and more clinically meaningful compared than the reductions seen in the SMBG group (Figure 2A). The largest HbA1c reduction was seen in older rtCGM users, age ≥50 years (Table 2).

Figure 2.

Changes in HbA1c from baseline: CGM vs SMBG (A) and CGM with MDI versus CGM with CSII (B).

Table 2.

Changes in HbA1c in rtCGM Group by Age.

Age range	CGM group					SMBG group					Diff of diff
Age range	N	Baseline	Measurement	Change	P value	N	Baseline	Measurement	Change	P value	Diff of diff	P value
≤19	23	8.65	8.64	−0.01	.657	274	8.80	8.86	0.06	.769	−0.07	.514
20-29	24	7.88	7.62	−0.26	.719	450	8.32	8.09	−0.23	.036	−0.03	.813
30-39	33	7.90	7.35	−0.55	.155	635	8.20	7.93	−0.27	.006	−0.28	.293
40-49	44	8.07	7.68	−0.39	.151	1270	8.39	8.13	−0.26	.000	−0.13	.452
50-59	51	8.74	8.04	−0.70	.049	2154	8.30	8.14	−0.16	<.001	−0.54	.037
≥60	12	8.35	7.58	−0.77	.021	1497	8.13	7.87	−0.26	<.001	−0.51	.210

Comparison of change in HbA1c according to insulin administration method showed a clinically significant HbA1c reduction with rtCGM+MDI but not rtCGM-CSII treatment; however, the between-group difference was not statistically significant (P = .06) (Figure 2B).

Secondary Analysis

The secondary analyses included a dataset of 22 925 patients from which 1130 were case matched (rtCGM, n = 565; SMBG, n = 565). No significant between-group differences in medical/pharmacy costs, gender, age, region, or comorbidities were observed. Patient demographics before and after propensity score matching are presented in Table 1.

Health Care Service Utilization

The number of all-cause inpatient admissions among rtCGM patients was significantly lower compared with SMBG patients with notably lower associated costs PMPM (Figure 3A). The number of diabetes-specific inpatient admissions and costs were also notably lower. The number of inpatient admissions coded for DKA among SMBG patients was more than double the number reported for rtCGM patients during the measurement period: 36 versus 16, P = .0675

Figure 3.

Utilization of health services and costs: Inpatient admissions (A) and ER admissions (B).

The number of all-cause ER visits was notably lower among rtCGM patients versus SMBG patients (P = .303) with associated lower costs PMPM (P = .491) (Figure 3B). The number of diabetes-specific ER admissions was also lower among rtCGM patients versus SMBG patients but with higher associated costs.

The number of ER visits coded for DKA among SMBG patients was more than four times higher than reported for rtCGM patients during the measurement period: 17 versus 4, P = .0318. Similar differences between rtCGM versus SMBG patients were also seen in the number of ER visits coded for hypoglycemia: 2 versus 7, P = .353

Discussion

Several recent studies have shown that use of rtCGM in conjunction with CSII treatment reduces HbA1c levels and occurrence severe of hypoglycemia and/or duration of hypoglycemia in pediatric and adult T1D populations.^7,14,18-21 However, recent evidence suggests that rtCGM use in conjunction with MDI treatment may provide the similar glycemic benefits without the additional costs associated with CSII therapy.²²

Our study found that use of rtCGM was associated with reduced HbA1c and utilization of health services compared with SMBG use regardless of insulin delivery method. In addition, rtCGM use was associated with notably fewer inpatient admissions and ER visits coded for DKA and hypoglycemia, which can have long-term effects on patient adherence. For example, as patients quickly learn that hypoglycemic episodes are physically aversive, potentially dangerous and a source of possible social embarrassment, they become fearful of hypoglycemia.²⁶ In many cases, this fear becomes a key obstacle to intensifying therapy and/or adhering to prescribed insulin regimens,^27,28 which often leads to poor metabolic control and subsequent health outcomes.²⁹

A 2016 report from the T1D Exchange registry showed that among CGM users there was no significant difference in HbA1c levels between CSII and MDI patients.³⁰ Our findings provide further evidence supporting rtCGM use regardless of insulin delivery method; however, the greater HbA1c improvements seen in the rtCGM-MDI group versus rtCGM-CSII add further support to the willingness of MDI-treated patients to utilize rtCGM data in their self-management. In a recent survey, Pettus and Edelman² found that rtCGM-MDI and rtCGM-CSII patients made similar changes in their correction and prandial insulin dosages based on their rtCGM data. Because rtCGM-MDI patients experienced greater HbA1c reductions, but without the added expense of CSII therapy, the cost savings may become an important factor when selecting an insulin administration method (MDI vs CSII).

We recognize that the higher baseline HbA1c levels among rtCGM+MDI users likely contributed to the more notable HbA1c reductions among rtCGM+MDI users. Therefore, although a more reasonable characterization of our findings would be that rtCGM use with either MDI and CSII confers similar glycemic benefits, rtCGM+MDI treatment may be considered more “cost effective” on the basis of this 1-year analysis as it eliminates the initial and ongoing costs of insulin pump use.

A key strength of this study is our utilization of propensity score analysis, which allowed us to identify equivalent comparison groups,³¹ thereby facilitating a true comparison between study groups. Another strength of our study design was use of longitudinal analysis used in a real life setting. This approach allowed us to measure the occurrence of change in SMBG acquisition behaviors and subsequent outcomes at the individual level, providing the opportunity to observe individual patterns of change.³²

A key limitation of the study is the small sample size of rtCGM users with pre and post HbA1c test values; a larger sample size would have provided a more robust assessment of the impact of rtCGM use on glycemic control and health service utilization. Another limitation was the design of the study. It is well known that retrospective analyses inherently include confounding variables, which may go unrecognized because of inadequate knowledge of how they interrelate with the outcomes. Although we were able to show associations between treatment modalities and outcomes, we cannot establish causal relationships. In addition, the Optum data set provided no information regarding the socioeconomic, educational characteristics or participation in a formal diabetes self-management education program, all of which could have affected outcomes.

Conclusions

Use of rtCGM in T1D patients facilitates greater improvements in HbA1c and reduced health care system utilization compared with traditional SMBG use regardless of insulin administration method. Treatment with rtCGM in conjunction with MDI confers similar or greater glycemic benefits without the additional costs associated with CSII therapy.

Footnotes

Acknowledgements

The authors wish to thank Stephanie MacLeod, Shirley Musich, Hungching Chan, Mark Holmberg, and Kevin Hawkins for their assistance in conducting the analysis.

Abbreviations

CSII, continuous subcutaneous insulin infusion; DKA, diabetic ketoacidosis; ER, emergency room; MDI, multiple daily insulin injections; PMPM, per-member per-month; rtCGM, real-time continuous glucose monitoring; SD, standard deviation; SMBG, self-monitoring of blood glucose; T1D, type 1 diabetes.

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: CGP has received consulting fees from Dexcom, Inc, Insulet Corporation, Roche Diabetes Care, and Sanofi. CG is an employee of Dexcom, Inc. JS has received consulting fees from Dexcom, Inc.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Funding for this study was provided by Dexcom, Inc, San Diego, CA 92121, USA.

References

Bailey

Grunberger

Bode

. American Association of Clinical Endocrinologists and American College of Endocrinology 2016 outpatient glucose monitoring consensus statement. Endocr Pract. 2016;22:231-261.

Pettus

Edelman

SV.

Use of glucose rate of change arrows to adjust insulin therapy among individuals with type 1 diabetes who use continuous glucose monitoring. Diabetes Technol Ther. 2016;18(suppl 20):S234-S242.

Pettus

Price

Edelman

SV.

How patients with type 1 diabetes translate continuous glucose monitoring data into diabetes management decisions. Endocr Pract. 2015;21:613-620.

Kovatchev

Patek

Ortiz

Breton

MD.

Assessing sensor accuracy for non-adjunct use of continuous glucose monitoring. Diabetes Technol Ther. 2015;17:177-186.

Vettoretti

Facchinetti

Sparacino

Cobelli

Simulation of patient decision-making to test safety and effectiveness of CGM sensor for non-adjunctive use in type 1 diabetes: SMBG vs CGM-based therapies. Paper presented at: 15th Annual Diabetes Technology Meeting; 22–24 October 2015; Bethesda, MD.

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;33:17-22.

Battelino

Conget

Olsen

. The use and efficacy of continuous glucose monitoring in type 1 diabetes treated with insulin pump therapy: a randomised controlled trial. Diabetologia. 2012;55:3155-3162.

Battelino

Phillip

Bratina

Nimri

Oskarsson

Bolinder

Effect of continuous glucose monitoring on hypoglycemia in type 1 diabetes. Diabetes Care. 2011;34:795-800.

Frontino

Bonfanti

Scaramuzza

. Sensor-augmented pump therapy in very young children with type 1 diabetes: an efficacy and feasibility observational study. Diabetes Technol Ther. 2012;14:762-764.

10.

New

Ajjan

Pfeiffer

Freckmann

Continuous glucose monitoring in people with diabetes: the randomized controlled Glucose Level Awareness in Diabetes Study (GLADIS). Diabet Med. 2015;32:609-617.

11.

Raccah

Sulmont

Reznik

. Incremental value of continuous glucose monitoring when starting pump therapy in patients with poorly controlled type 1 diabetes: the RealTrend study. Diabetes Care. 2009;32:2245-2250.

12.

Buse

Kudva

Battelino

Davis

Shin

Welsh

JB.

Effects of sensor-augmented pump therapy on glycemic variability in well-controlled type 1 diabetes in the STAR 3 study. Diabetes Technol Ther. 2012;14:644-647.

13.

Hermanns

Schumann

Kulzer

Haak

The impact of continuous glucose monitoring on low interstitial glucose values and low blood glucose values assessed by point-of-care blood glucose meters: results of a crossover trial. J Diabetes Sci Technol. 2014;8:516-522.

14.

Hommel

Olsen

Battelino

. Impact of continuous glucose monitoring on quality of life, treatment satisfaction, and use of medical care resources: analyses from the SWITCH study. Acta Diabetol. 2014;51:845-851.

15.

Argento

Nakamura

Personal real-time continuous glucose monitoring in patients 65 years and older. Endocr Pract. 2014;20:1297-1302.

16.

Juvenile Diabetes Research Foundation Continuous Glucose Monitoring Study Group, Beck

Buckingham

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

17.

Chamberlain

Dopita

Gilgen

Neuman

Impact of frequent and persistent use of continuous glucose monitoring (CGM) on hypoglycemia fear, frequency of emergency medical treatment, and SMBG frequency after one year. J Diabetes Sci Technol. 2015;10(2):383-388.

18.

Picard

Hanaire

Baillot-Rudoni

. Evaluation of the adherence to continuous glucose monitoring in the management of type 1 diabetes patients on sensor-augmented pump therapy: the SENLOCOR study. Diabetes Technol Ther. 2016;18(3):127-135.

19.

Choudhary

Olsen

Conget

Welsh

Vorrink

Shin

JJ.

Hypoglycemia prevention and user acceptance of an insulin pump system with predictive low glucose management. Diabetes Technol Ther. 2016;18(5):288-291.

20.

Esvant

Lefebvre

Campillo-Gimenez

. A mobile application guiding patients with type 1 diabetes using sensor-augmented insulin pump therapy. J Diabetes Sci Technol. 2016;10:985-986.

21.

Triolo

Maahs

Pyle

. Effects of frequency of sensor-augmented pump use on hba1c and c-peptide levels in the first year of type 1 diabetes. Diabetes Care. 2016;dc152201. http://dx.doi.org/10.2337/dc15-2201.

22.

Tumminia

Crimi

Sciacca

. Efficacy of real-time continuous glucose monitoring on glycaemic control and glucose variability in type 1 diabetic patients treated with either insulin pumps or multiple insulin injection therapy: a randomized controlled crossover trial. Diabetes Metab Res Rev. 2015;31:61-68.

23.

Roze

Smith-Palmer

Valentine

. Cost-effectiveness of sensor-augmented pump therapy with low glucose suspend versus standard insulin pump therapy in two different patient populations with type 1 diabetes in France. Diabetes Technol Ther. 2016;18:75-84.

24.

Roze

Smith-Palmer

Valentine

. Long-term health economic benefits of sensor-augmented pump therapy vs continuous subcutaneous insulin infusion alone in type 1 diabetes: a UK perspective. J Med Econ. 2016;19:236-242.

25.

Quan

Couris

. Updating and validating the Charlson comorbidity index and score for risk adjustment in hospital discharge abstracts using data from 6 countries. Am J Epidemiol. 2011;173:676-682.

26.

Brouhard

. Hypoglycemia. In: Diabetes Mellitus in Children and Adolescents. Philadelphia, PA: W.B. Saunders; 1987:169-178.

27.

Cryer

PE.

Hypoglycemia: still the limiting factor in the glycemic management of diabetes. Endocr Pract. 2008;14:750-756.

28.

Riddle MC: The underuse of insulin therapy in North America. Diabetes Metab Res Rev. 2002;18(suppl 3):S42-S49.

29.

Wild

von Maltzahn

Brohan

Christensen

Clauson

Gonder-Frederick

A critical review of the literature on fear of hypoglycemia in diabetes: Implications for diabetes management and patient education. Patient Educ Couns. 2007;68:10-15.

30.

Foster

Miller

Tamborlane

Bergenstal

Beck

, Network TDEC. Continuous glucose monitoring in patients with type 1 diabetes using insulin injections. Diabetes Care. 2016;39:e81-82.

31.

Rosenbaum

Rubin

DB.

The central role of the propensity score in observational studies for causal effects. Biometrika. 1983;70:41-55.

32.

Toh

Hernan

MA.

Causal inference from longitudinal studies with baseline randomization. Int J Biostat. 2008;4(1):article 22. Available at: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2835458/pdf/ijb1117.pdf. Accessed May 8, 2016.

Continuous Glucose Monitoring Use in Type 1 Diabetes: Longitudinal Analysis Demonstrates Meaningful Improvements in HbA1c and Reductions in Health Care Utilization

Abstract

Background:

Methods:

Results:

Conclusions:

Keywords

Methods

Design

Analysis

Results

Primary Analysis

Secondary Analysis

Health Care Service Utilization

Discussion

Conclusions

Footnotes

Acknowledgements

Abbreviations

Declaration of Conflicting Interests

Funding

References