FRONTIER—FreeStyle Libre System Use in Ontario Among People with Diabetes Mellitus in the IC/ES Database—Evidence from Real-World Practice: Patients Using Intensive Insulin

Abstract

Background:

Diabetes mellitus is associated with significant health care resource utilization (HCRU), partly due to acute complications, including diabetic ketoacidosis (DKA) and hypoglycemia.

Aim:

To investigate glycated hemoglobin (HbA1c) levels and HCRU before and after adoption of FreeStyle Libre Systems (FSL) in people with diabetes on multiple daily injections of insulin (MDI).

Methods:

This retrospective longitudinal study used administrative health data in Ontario, Canada, housed at IC/ES. The cohort comprised people with diabetes on MDI with a first FSL claim between September 16, 2019, and August 31, 2020 (index date), who remained on FSL for 24 months. HCRU (emergency department [ED] visits and hospitalization) was measured for 12 months before the index date and the last 12 months of follow-up. HbA1c data were taken from the last tests in each period.

Results:

Mean HbA1c was statistically significantly reduced after FSL among people with type 1 diabetes mellitus (T1DM; n = 10,510; age <25 years, −0.8%; 25–65 years, −0.5%; >65 years, −0.1%; all P < 0.0001) or type 2 diabetes mellitus (T2DM; n = 12,668; age ≤65 years, −0.6%; >65 years, −0.3%; both P < 0.0001). Overall HCRU was statistically significantly reduced in the T1DM subgroups aged <25 and 25–65 years (ED visits only) and both T2DM age subgroups, with some subgroups having statistically significant reductions in DKA- or hypoglycemia-associated HCRU.

Conclusions:

Among people with T1DM or T2DM on MDI, HbA1c was statistically significantly reduced after FSL, with statistically significant reductions in HCRU in some subgroups.

Introduction

Diabetes mellitus is associated with significant and rising health care resource utilization (HCRU).¹ In 2019, Canadians spent an estimated $30 billion (Canadian dollars) on diabetes-related health care expenses, a steep increase from the $14 billion spent in 2008.² The cost of diabetes to the Canadian health care system is projected to exceed $39 billion by 2028.³ In addition to long-term complications, acute complications resulting from suboptimal glycemic management are important drivers of the economic burden of diabetes; these acute complications include diabetic ketoacidosis (DKA) and hypoglycemia, which can require emergency department (ED) visits or hospitalization.^4,5

Effective self-management is imperative to maintaining stable glucose levels,⁶ and reduces the risk of acute complications and acute events among people with type 1 diabetes mellitus (T1DM) or type 2 diabetes mellitus (T2DM).^7,8 However, glycemic self-management is not always optimal among Canadians with diabetes, particularly in younger age groups.^9,10

Sensor-based glucose monitoring systems may support users to improve their self-management and related outcomes by supplying key insights into glucose readings, glucose variability, and trends.¹¹

FreeStyle Libre Systems (FSL) are factory-calibrated, sensor-based devices that allow people living with diabetes to monitor glucose levels, variability, and trends.¹¹ Trials and real-world studies conducted outside Canada demonstrated lowered glycated hemoglobin (HbA1c) levels and increased time spent in the recommended glycemic range in people with T1DM or T2DM using FSL.^12

–15 In addition, several cohort studies have shown use of FSL to be associated with reductions in HCRU due to acute complications.^7,16,17 However, there is a lack of population-based data on the effectiveness of FSL in Canada.

Guidelines in the United States currently recommend the use of sensor-based glucose monitoring devices for people living with diabetes who are using insulin or who have a history of problematic hypoglycemia.^6,18,19 In Canada, sensor-based glucose monitoring devices are recommended for individuals with T1DM or T2DM treated with intensive insulin.²⁰ However, public coverage of FSL varies across Canadian provinces and territories, with only Ontario having full coverage, through the Ontario Drug Benefit (ODB), for all people with diabetes using any insulin regimen.^21,22 In Alberta, Manitoba, Quebec, and Nova Scotia, FSL is covered for people with T1DM or T2DM using intensive insulin, but Yukon provides coverage only for people with T1DM. In other provinces and territories, FSL is either not covered for adults or is covered on a case-by-case basis.²¹

The present study assesses the effect of FSL on HbA1c and HCRU among adults with T1DM or T2DM on intensive insulin in Canada using real-world health care data from Ontario. Ontario is a large province with an especially high prevalence of diabetes (11%, compared with the national average of 9.4%) and comprehensive administrative health databases, which allows a large population-based study to be conducted.^22
–24

Objectives

This study aimed to investigate HbA1c levels and HCRU before and after adoption of FSL in people with diabetes on intensive insulin therapy, defined as the use of multiple daily injections of insulin (MDI) or insulin pumps.

Methods

Data source

This repeated cross-sectional study used routinely collected administrative health data in Ontario, Canada, housed at IC/ES (formerly known as the Institute for Clinical Evaluative Sciences). The IC/ES data repository includes coded and linkable administrative health data for 13 million people (including up to 1.24 million with diabetes, depending on the criteria used to identify cases).^24,25 Specifically, information is contained on ED visits, hospitalizations, physician visits, and prescription drug use for all residents of Ontario who are covered by the Ontario Health Insurance Plan (OHIP) and ODB.

Identification of diabetes treatment and FSL claims

For inclusion, individuals were required to have a confirmed diagnosis of diabetes in the Ontario Diabetes Database, which contains records of people with diabetes resident in Ontario, identified through a validated algorithm,²⁶ or at least one diagnosis or medication code for diabetes (International Classification of Diseases, Clinical Modification, 250.xx; Anatomical Therapeutic Chemical class, A10), as previously validated.²⁷ For inclusion in the final analysis, patients had to be receiving intensive insulin (MDI or insulin pumps; pump use was identified using Ontario Assistive Devices Program data).

Study population

The analysis cohort comprised people with diabetes using intensive insulin who had a first FSL claim between September 16, 2019, and August 31, 2020 (index date), and remained active on FSL for 24 months of follow-up after the index date, with a ≥70% medication possession ratio (used to measure adherence, as in previous studies²⁸; Fig. 1).

FIG. 1.

Study schematic. HbA1c, glycated hemoglobin; HCRU, health care resource utilization; ODB, Ontario Drug Benefit.

The cohort was split into subgroups aged ≤65 or >65 years at the index date (the >65 years age threshold ensures ODB coverage throughout the 12-month preindex period). For people with T1DM, the subgroup aged ≤65 years was further divided into ages <25 and 25–65 years. This was done because people in the IC/ES data repository aged <25 years are typically covered by the OHIP and may have different socioeconomic characteristics from those aged 25–65 years, who are more likely to be covered by ODB. For people with T2DM, this subdivision was not applied due to the small number of T2DM cases expected among people aged <25 years.

People with diabetes using insulin pumps were not split into T1DM and T2DM subgroups because they could not be differentiated using IC/ES data; nonetheless, the number of people with T2DM using pumps is expected to be small.

Treatment history

The proportions of patients in each group who had a history of treatment for common diabetes complications—cardiovascular disease, diabetic retinopathy, hyperlipidemia, hypertension, neuropathy, peripheral vascular disease, and stroke—in the last 12 months before the index date were determined from the IC/ES database. Treatment history in the database reflects people with diabetes having attended an ED or inpatient visit to which the International Classification of Diseases (10th revision; ICD-10) code for the relevant condition was attached, or a regular doctor’s visit coded specifically for that condition. It is therefore likely to represent people with severe conditions and is not a measure of the prevalence of individual comorbidities (e.g., a diabetes review appointment for a patient with hypertension would not typically have a hypertension code associated with it).

HbA1c data

HbA1c data were taken from the most recent laboratory test before the index date and from the latest test in the follow-up period.

Measurement of HCRU

For each included patient, HCRU was measured for 12 months before the index date and the last 12 months of the 24-month follow-up period. Annualized incidence rates were calculated for all-cause ED visits and hospitalizations during each period, as well as for ED visits and hospitalizations specifically associated with (1) DKA and (2) hypoglycemia.

Resource intensity weight (RIW) data were used to compare the level of HCRU for each ED visit and hospitalization. In Canada, RIW is determined by the Canadian Institute for Health Information using hospital discharge data and is designed to allow the estimation of a nationally comparable cost for each hospital stay, including the costs of administration, staff, supplies, and equipment.²⁹ The higher a person’s RIW, the more hospital resources they consumed during their stay.³⁰ RIW data have previously been used in studies of the health care burden of diabetes to estimate the costs associated with hospital admissions.^29,31

Statistical analysis

The analysis was conducted using SAS Enterprise Guide v8.3 (SAS Institute, Cary, North Carolina, USA). Paired t-tests were conducted to compare HbA1c and HCRU before and after FSL acquisition among individuals with available data for both periods. P values of <0.025 were considered to be statistically significant; this threshold was selected to balance false-positive and false-negative findings.

Results

Analysis cohort

In total, 24,074 people with diabetes who were using intensive insulin therapy had a first FSL claim during the selection period (T1DM on MDI, n = 10,510; T2DM on MDI, n = 12,668; insulin pump users, n = 896; Table 1).

Table 1.

Patient Characteristics

Characteristic	T1DM MDI			T2DM MDI		Insulin pump^a
Characteristic	Age <25 years	Age 25–65 years	Age >65 years^b	Age ≤65 years	Age >65 years^b	Age <25 years	Age 25–65 years	Age >65 years^b
Sample size	703	3,546	6,261	3,322	9,346	301	328	267
Age at index date, years
Mean (SD)	15.381 (5.575)	53.191 (10.753)	73.952 (6.089)	56.589 (10.433)	74.014 (6.127)	18.033 (4.387)	48.476 (13.781)	70.772 (4.155)
Median (Q1–Q3)	16 (11–20)	56 (47–62)	73 (69–78)	60 (53–64)	73 (69–78)	19 (15–22)	51 (37–62)	70 (68–73)
Min–Max	2–24	25–65	66–104	7–65	66–100	6–24	25–65	66–87
Male, n (%)	377 (53.6%)	1,878 (53.0%)	3,254 (52.0%)	1,811 (54.5%)	5,263 (56.3%)	172 (57.1%)	144 (43.9%)	126 (47.2%)
Income quintile, n (%)
1	231 (32.9%)	1,447 (40.8%)	1,631 (26.1%)	1,228 (37.0%)	2,224 (23.8%)	63 (20.9%)	87 (26.5%)	50 (18.7%)
2	135 (19.2%)	768 (21.7%)	1,374 (21.9%)	750 (22.6%)	2,106 (22.5%)	59 (19.6%)	71 (21.6%)	39 (14.6%)
3	142 (20.2%)	568 (16.0%)	1,210 (19.3%)	588 (17.7%)	1,859 (19.9%)	56 (18.6%)	61 (18.6%)	58 (21.7%)
4	124 (17.6%)	437 (12.3%)	1,117 (17.8%)	431 (13.0%)	1,686 (18.0%)	63 (20.9%)	55 (16.8%)	52 (19.5%)
5	66–70 ^c	316 (8.9%)	912 (14.6%)	317 (9.5%)	1,454 (15.6%)	55–59 ^c	54 (16.5%)	68 (25.5%)
Missing	1–5 ^c	10 (0.3%)	17 (0.3%)	8 (0.2%)	17 (0.2%)	1–5 ^c	0 (0.0%)	0 (0.0%)
Rural residence, n (%)^d	59 (8.4%)	436 (12.3%)	828 (13.2%)	340 (10.2%)	1,148 (12.3%)	43 (14.3%)	50 (15.2%)	53 (19.9%)
Duration in ODD, years
Mean (SD)	5.726 (5.157)	18.490 (7.807)	22.475 (6.301)	14.764 (7.752)	19.836 (7.109)	9.574 (4.813)	21.699 (7.563)	23.932 (6.530)
Median (Q1–Q3)	5 (1–9)	19 (13–26)	25 (19–28)	15 (9–21)	21 (15–26)	9 (6–13)	24 (16–28)	27 (21–28)
Treatment history—12 months before index date, n (%)^e
Cardiovascular disease	11 (1.6%)	832 (23.5%)	2,367 (37.8%)	695 (20.9%)	2,974 (31.8%)	1–5^c	46 (14.0%)	58 (21.7%)
Diabetic retinopathy	6 (0.9%)	386 (10.9%)	935 (14.9%)	323 (9.7%)	1,055 (11.3%)	1–5 ^c	42 (12.8%)	42 (15.7%)
Hyperlipidemia	1–5^c	159 (4.5%)	266 (4.2%)	176 (5.3%)	383 (4.1%)	1–5^c	11–15^c	7 (2.6%)
Hypertension	11 (1.6%)	882 (24.9%)	2,027 (32.4%)	830 (25.0%)	2,771 (29.6%)	9 (3.0%)	59 (18.0%)	61 (22.8%)
Neuropathy	6 (0.9%)	354 (10.0%)	298 (4.8%)	159 (4.8%)	265 (2.8%)	1–5^c	16 (4.9%)	8 (3.0%)
Peripheral vascular disease	0 (0.0%)	232 (6.5%)	424 (6.8%)	138 (4.2%)	474 (5.1%)	1–5^c	1–5^c	16 (6.0%)
Stroke	1–5^c	148 (4.2%)	302 (4.8%)	88 (2.6%)	339 (3.6%)	0 (0.0%)	6 (1.8%)	7 (2.6%)

Because of the small number of people with T2DM expected to be using insulin pumps, insulin pump use was not analyzed by diabetes type.

The >65-year-age threshold ensures ODB coverage during the 12-month preindex period.

To protect patient confidentiality, data for any characteristics with patient count of 1–5 could not be provided by IC/ES.

Rural residence is defined as living in a community with ≤10,000 people.

Treatment history reflects individuals having attended an ED or inpatient visit to which the ICD-10 code for the relevant condition was attached, or a regular doctor’s visit coded specifically for that condition.

ED, emergency department; ICD-10, International Classification of Diseases, 10th revision; MDI, multiple daily injections of insulin; ODB, Ontario Drug Benefit; ODD, Ontario Diabetes Database; Q, quartile; SD, standard deviation; T1DM, type 1 diabetes mellitus; T2DM, type 2 diabetes mellitus.

The most common conditions for which individuals had a treatment history in the 12 months before the index date were cardiovascular disease and hypertension (Table 1). Patients aged <25 years were less likely to have received treatment in the last 12 months than the older cohorts. Treatment history was generally comparable across treatment groups, although a history of cardiovascular disease appeared to be less common in the insulin pump groups than in the MDI groups, and a history of neuropathy appeared to be more common in the T1DM MDI group than in the other cohorts.

HbA1c level

The mean (standard deviation [SD]) HbA1c level at the last test before the index date was 9.8% (2.4%) among people with T1DM on MDI aged <25 years, 8.9% (1.9%) among those aged 25–65 years, and 8.1% (1.4%) among those aged >65 years (Fig. 2A). After FSL acquisition, mean HbA1c was statistically significantly lower in all three age subgroups (age <25 years, −0.8% [SD, 2.5%]; age 25–65 years, −0.5% [1.7%]; age >65 years, −0.1% [1.3%]; all P < 0.0001; Fig. 2A).

FIG. 2.

Mean HbA1c among people with (A) T1DM or (B) T2DM who were using MDI, before and after acquisition of FSL. Data and P values are for individuals with available data for both pre- and post-FSL periods. Values are shown as mean (SD). FSL, FreeStyle Libre; MDI, multiple daily injections of insulin; SD, standard deviation; T1DM, type 1 diabetes mellitus; T2DM, type 2 diabetes mellitus.

Among people with T2DM on MDI, mean HbA1c before acquisition of FSL was 8.6% (SD, 1.7%) among those aged ≤65 years, and 8.1% (1.4%) in the age >65 subgroup (Fig. 2B). After FSL acquisition, statistically significant mean reductions were seen in both age subgroups (age ≤65 years, −0.6% [SD, 1.7%]; age >65 years, −0.3% [1.4%]; both P < 0.0001; Fig. 2B).

For people with diabetes using insulin pumps, a statistically significant reduction in HbA1c was seen only in the subgroup aged <25 years (from 8.9% to 8.5%; P = 0.0002; Supplementary Table S1).

The proportions of individuals who did not have HbA1c tests in the periods of interest are shown in Supplementary Table S2. For all treatments, missing data were more common in the younger cohorts.

Health care resource utilization

HCRU before and after FSL acquisition is shown in Table 2.

Table 2.

HCRU Before and After Acquisition of FreeStyle Libre Among People with Diabetes Using Multiple Daily Injections of Insulin

Outcome and subgroup	Before FSL	After FSL	Percentage change	P value
T1DM
ED visits per patient per year
Overall
Age < 25 years	1.148 (2.291)	0.808 (1.917)	−29.6%	<0.0001
Age 25–65 years	2.098 (3.872)	1.873 (3.760)	−10.8%	0.0002
Age >65 years	1.215 (2.042)	1.170 (2.178)	−3.7%	0.129
With DKA
Age < 25 years	0.253 (0.851)	0.176 (0.682)	−30.4%	0.014
Age 25–65 years	0.095 (0.503)	0.079 (0.539)	−16.8%	0.083
Age >65 years	0.014 (0.132)	0.010 (0.117)	−21.4%	0.134
With hypoglycemia
Age < 25 years	0.121 (0.478)	0.060 (0.328)	−50.4%	0.002
Age 25–65 years	0.136 (0.537)	0.108 (0.489)	−20.6%	0.008
Age >65 years	0.056 (0.268)	0.046 (0.269)	−16.1%	0.036
Hospitalization per patient per year
Overall
Age < 25 years	0.384 (0.951)	0.232 (0.873)	−39.6%	<0.0001
Age 25–65 years	0.497 (1.194)	0.488 (1.315)	−1.8%	0.679
Age >65 years	0.337 (0.770)	0.344 (0.823)	+2.1%	0.589
With DKA
Age < 25 years	0.256 (0.813)	0.168 (0.665)	−34.4%	0.009
Age 25–65 years	0.104 (0.504)	0.091 (0.629)	−12.5%	0.213
Age >65 years	0.016 (0.150)	0.015 (0.157)	−12.5%	0.547
With hypoglycemia
Age < 25 years	0.060 (0.281)	0.046 (0.289)	−23.3%	0.308
Age 25–65 years	0.133 (0.508)	0.131 (0.552)	−1.5%	0.854
Age >65 years	0.063 (0.282)	0.054 (0.266)	−14.3%	0.053
T2DM
ED visits per patient per year
Overall
Age ≤65 years	0.836 (2.118)	0.636 (1.576)	−23.9%	<0.0001
Age >65 years	0.696 (1.371)	0.610 (1.307)	−12.4%	<0.0001
With DKA
Age ≤65 years	0.015 (0.129)	0.005 (0.069)	−66.7%	<0.0001
Age >65 years	0.006 (0.083)	0.002 (0.094)	−66.7%	0.002
With hypoglycemia
Age ≤65 years	0.023 (0.158)	0.018 (0.143)	−21.7%	0.170
Age >65 years	0.025 (0.179)	0.013 (0.122)	−48.0%	<0.0001
Hospitalization per patient per year
Overall
Age ≤65 years	0.213 (0.602)	0.167 (0.634)	−21.1%	0.0004
Age >65 years	0.238 (0.629)	0.209 (0.635)	−11.8%	0.0007
With DKA
Age ≤65 years	0.016 (0.136)	0.005 (0.071)	−68.8%	<0.0001
Age >65 years	0.009 (0.103)	0.003 (0.096)	−66.7%	<0.0001
With hypoglycemia
Age ≤65 years	0.048 (0.251)	0.025 (0.174)	−47.9%	<0.0001
Age >65 years	0.040 (0.224)	0.025 (0.180)	−37.5%	<0.0001

Data are presented as univariate mean (standard deviation). Data and P values are for individuals with available data for both pre- and post-FSL periods. Statistically significant P values are shown in bold.

DKA, diabetic ketoacidosis.

Among people with T1DM on MDI, the rates of ED visits and hospitalization, overall, with DKA or with hypoglycemia, were numerically lower after acquisition of FSL compared with the preindex period, except for the overall hospitalization among those aged >65 years. The reductions in overall ED visits were statistically significant in the subgroups aged <25 years (percentage change, −29.6%) and 25–65 years (−10.8%), but not in the subgroup aged >65 years (Table 2). In addition, there were statistically significant reductions in ED visits with DKA among people aged <25 years (−30.4%), and in ED visits with hypoglycemia in the subgroups aged <25 years (−50.4%) and 25–65 years (−20.6%; Table 2). Overall hospitalization (−39.6%) and hospitalization with DKA (−34.4%) were statistically significantly reduced in the subgroup aged <25 years (Table 2). Other changes were not statistically significant.

For people with T2DM on MDI, ED visits overall (age ≤65 years, −23.9%; age >65 years, −12.4%), with DKA (−66.7% in both age subgroups), and with hypoglycemia (age ≤65 years, −21.7%; age >65 years, −48.0%) were statistically significantly reduced after acquisition of FSL in both age subgroups, with the exception of the reduction in ED visits with hypoglycemia among those aged ≤65 years (Table 2). Similarly, hospitalization overall (age ≤65 years, −21.1%; age >65 years, −11.8%), with DKA (age ≤65 years, −68.8%; age >65 years, −66.7%), and with hypoglycemia (age ≤65 years, −47.9%; age >65 years, −37.5%) were statistically significantly reduced in both age subgroups (Table 2).

The mean RIW (i.e., measure of resource use per event) for each ED visit and hospitalization showed no statistically significant changes in any group after initiation of FSL (Supplementary Table S3).

Among people with diabetes using insulin pumps, no statistically significant changes in any HCRU measures were seen (Supplementary Table S1).

Discussion

Diabetes remains a major health concern in Ontario, with a prevalence higher than the Canadian average.²² The prevalence of diabetes (T1DM and T2DM combined) is projected to increase by 25% from 2023 to 2033.²²

The results of this study showed that, among people with diabetes in Ontario using MDI, adoption of FSL improved HbA1c and reduced overall HCRU. Mean HbA1c was statistically significantly reduced both among people with T1DM and T2DM, and across all age groups. Also, in most subgroups, the mean rates of ED visits and hospitalization (overall, with DKA, and with hypoglycemia) were lower after FSL acquisition than in the preindex period. By contrast, among people with diabetes using insulin pumps, the reduction in HbA1c was statistically significant only in the youngest subgroup, and no significant changes in HCRU were seen.

This study addresses the limited real-world evidence for the effectiveness of FSL in Canada, demonstrating improvements in HbA1c and HCRU in a large, ethnically diverse population.²²

The benefits of sensor-based glucose monitoring for people living with T1DM and T2DM are well established, with extensive evidence available from clinical trials and real-world studies. A meta-analysis of 25 studies (clinical trials and real-world observational studies) showed that the use of FSL was associated with a significant and sustained reduction of HbA1c in adults with T1DM or T2DM (overall reduction, −0.6%, compared with −0.1% to −0.8% seen across groups in the present study).¹³ A retrospective cohort study (RELIEF) conducted using the French national health claims database showed reductions in hospitalizations for DKA (−55.0%) and hypoglycemia (−6.4%) in the 12 months following FSL initiation by people with T1DM or T2DM (n = 74,011).³² Reductions in acute complications (−40% at 24 months) were also observed in a subgroup analysis of people aged ≥65 years with T2DM on intensive insulin therapy.⁷ In the present analysis, broadly similar reductions in hospitalization for DKA were seen overall (T1DM, −12.5% to −34.4%; T2DM, −66.7% to −68.8% [reductions were not statistically significant in T1DM groups aged ≥25 years, perhaps reflecting the lower baseline event rates in these groups than in the youngest subgroup]), with larger reductions in hospitalization for hypoglycemia in some groups (T1DM, −1.5% to −23.3%; T2DM, −37.5% to −47.9% [reductions were statistically significant only in T2DM groups]) compared with the French study.

Improvements in HbA1c and HCRU have also been described for other sensor-based monitoring devices. For example, in a retrospective study of people with T1DM or T2DM treated with MDI in the Kaiser Permanente Northern California health care system (mean baseline HbA1c was 8.2% in both T1DM and T2DM groups), initiation of a continuous glucose monitoring (CGM) device was associated with significant reductions in HbA1c compared with people not initiating CGM use (T1DM, −0.34%; T2DM, −0.56%).³³ In addition, both ED visits and hospitalization for hypoglycemia were numerically lower after CGM initiation.³³ Reductions in HCRU were also seen after initiation of a CGM device in a recent retrospective analysis of people with T2DM using intensive insulin in the Optum Clinformatics Data Mart database: the proportions of individuals with at least one diabetes ED visit or inpatient visit were reduced by 30.0% and 41.5%, respectively.³⁴

A further retrospective cohort study, using the IBM MarketScan Research Databases, found that people with T1DM or T2DM using intensive insulin experienced similar reductions in HbA1c after acquisition of FSL or another CGM device.³⁵ HbA1c reductions with FSL and another CGM device were −0.35% and −0.37%, respectively, among people with T1DM, and −0.73% and −0.79%, respectively, in the T2DM group, with no significant differences between devices in either population (mean baseline HbA1c levels were 8.5%–9.1% across groups).³⁵ In the same retrospective study, people with T1DM or T2DM experienced similar rates of acute diabetes events and all-cause hospitalizations after obtaining FSL or another CGM device.³⁵ In the present study, acquisition of FSL was associated with reductions in HbA1c and HCRU in almost all age and diabetes type subgroups on MDI. In both T1DM and T2DM MDI subgroups, the cohorts aged ≤65 years had higher mean HbA1c at baseline compared with those in the older groups, a finding that has also been observed in a previous study investigating age differences in HbA1c.³⁶ The younger age groups also experienced greater mean reductions in HbA1c after FSL acquisition than the older age groups. This difference may reflect the higher initial mean HbA1c level compared with the older age groups, but could also be a consequence of greater utilization of FSL monitoring in the younger cohort. It is also possible that this is a reflection of HbA1c goals being relaxed as patients age, as suggested by the Canadian guidelines for the management of diabetes in older people.^37,38

In this analysis, people initiating FSL during the study period had higher mean HbA1c levels than has been reported for the overall Ontario diabetes population. In 2019, a mean HbA1c level of 7.2% was reported across 1.2 million individuals with diabetes, with only 19% having an HbA1c level of >8.0%.³⁹ By contrast, only one of the subgroups included in the present study (age >65 years, using insulin pumps) had a mean HbA1c level below 8.0%. The reasons for this disparity are unclear. However, it is notable that because public funding for health care for those aged <65 years is limited to government-supported individuals who cannot work (e.g., those on Ontario Works or the Ontario Disability Support Program), the cohort aged ≤65 years reflects only a small proportion of people with diabetes in that age group and may therefore not be representative of the overall population. Even after the reductions seen following FSL acquisition, mean HbA1c levels, especially in the younger cohorts, still exceeded recommended glycemic targets. It has previously been suggested that diabetes is undertreated in some younger patients in Canada,⁹ and for many patients additional interventions are necessary.

Reductions in HCRU were statistically significant for people with T1DM on MDI only in the younger age groups. This may reflect the lower baseline rates of ED visits and hospitalizations due to DKA or hypoglycemia seen in people with T1DM aged >65 years than in younger groups. By comparison, statistically significant reductions in HCRU were observed for all people with T2DM on MDI (aged either ≤65 years or >65 years). The reason for this difference in effect between T1DM and T2DM in the older age group is unclear, but does not appear to be due to differences in baseline HCRU: for both ED visits and hospitalizations, including those due to DKA or hypoglycemia, the T2DM subgroup aged >65 years had lower baseline rates and higher percentage reductions than the T1DM subgroup aged >65 years.

Compared with the T1DM and T2DM MDI populations, patients using insulin pumps had lower mean HbA1c levels before FSL acquisition, which may have reduced the potential for significant reductions to occur. It is unclear why no statistically significant reductions in HCRU were seen in the insulin pump cohort. This analysis is limited by the small number of patients in the insulin pump groups and the lack of data as to how FSL was being used by these patients (e.g., whether there was any linkage between the devices). The study cohort comprised patients initiating FSL in 2019–2020, and their use of the technology may differ from patients prescribed FSL more recently, given the approval of hybrid closed-loop insulin delivery by Health Canada in 2024.

Although there were statistically significant decreases in both ED visits and hospitalizations, there was expectedly no statistically significant change in the relative cost of each health care encounter, as estimated by the RIW, before and after FSL initiation. In other words, the adoption of FSL did not alter the complexity or resource demands of ED visits or hospitalization among people living with diabetes. Since the number of ED visits and hospitalization events fell after FSL acquisition among people with T1DM or T2DM who were on MDI, this stable RIW suggests an overall reduction in HCRU costs.

This study confirms the benefits of FSL acquisition among people living with diabetes who are using MDI. Reductions in HbA1c and in HCRU due to DKA and hypoglycemia were seen among people with T2DM on MDI as well as for those with T1DM. Despite these benefits, however, FSL is not covered by public health care in some Canadian provinces and territories. If other provinces were to adopt the coverage policy used in Ontario, thereby broadening access to FSL, this might help alleviate the substantial burden of diabetes in Canada. Further studies in other parts of Canada would be valuable, although these are challenging due to the lack of an equivalent to the IC/ES data repository of linked health data in other provinces.

The main strength of this study is the use of large provincial health administrative databases that include the health records of all people who are eligible for publicly funded health care in Ontario, an ethnically diverse province with a high prevalence of diabetes.²² Coded identifiers enable the linkage of physician claims submitted to the OHIP, medical drug claims submitted to the ODB program, and discharge summaries for hospital stays and ED visits, as well as other data sources.²⁴ This enables a population-based study to be carried out with a comprehensive scope and a large sample, selected from among all people with diabetes in Ontario receiving publicly funded health care.

However, this study has several limitations. First, as described above, the cohort aged ≤65 years reflects only a small proportion of people with diabetes in that age group. The potentially unrepresentative nature of the cohort aged ≤65 years was partly addressed by analyzing people aged <25 years and those aged 25–65 years separately; these groups are mainly covered by OHIP+ and ODB, respectively. The OHIP+ program covers prescription medications for people aged <25 years, while the ODB (which mainly provides cover for people aged ≥65 years) covers people aged younger than 65 years only if they are living in long-term care or enrolled in specific support programs. A substantial proportion of individuals in the subgroups aged <25 years had missing HbA1c data; results in those subgroups should therefore be treated with caution. The use of age categorization may have masked important within-group heterogeneity. In the insulin pump group, it was not possible to distinguish between people with T1DM and those with T2DM. It was not possible to determine whether there was any linkage between FSL and insulin pumps. Given the date range of the study, people using the diabetes medications approved after August 31, 2020 (in particular, tirzepatide, which was not available in Canada at the time of the study), were not included. As the study was conducted using routinely collected administrative health data, no information was available on the reasons for initiating CGM, the education provided at the time of starting CGM, if any, the number of sensors used, the number of scans per day, or the total amount of glucose data collected. Similarly, no patient-reported outcomes data were collected. No formal multiplicity correction was conducted, with a P < 0.025 threshold selected to balance false-positive and false-negative findings. Lastly, only the acquisition of FSL, and not the extent to which people made use of the devices, could be determined. The possibility of nonadherence to FSL use means that the observed changes in HbA1c and HCRU may underestimate the effectiveness of FSL when used continuously. Further limitations of this analysis include the lack of a control group and the potential for confounding, particularly given the potential for period effects associated with the COVID-19 pandemic.

Conclusions

Among people with diabetes in Canada who were on MDI, but not for most outcomes among those using insulin pumps, HbA1c levels were reduced statistically significantly after initiation of FSL. Most measures of HCRU were reduced statistically significantly among people with T2DM, with statistically significant reductions also seen for some HCRU measures among people with T1DM. In general, larger reductions were seen in younger age groups using MDI, compared with older patients, particularly among those with T1DM. Benefits were seen for patients with T2DM on MDI, as well as for those with T1DM. The reductions in HCRU observed after FSL acquisition suggest that broadening access to FSL in Canada could help address the growing burden of diabetes.

Footnotes

Acknowledgments

Medical writing support was provided by Dr. Paul Overton (Beacon Medical Communications Ltd, Brighton, UK) in accordance with Good Publication Practice (GPP 2022) guidelines and was funded by Abbott.

Accurytics Ltd. (Richmond Hill, Ontario) supported in developing the study protocol and managing the study with IC/ES, which was funded by Abbott. The conclusions, opinions, and statements expressed herein are solely those of the authors and no endorsement is intended or should be inferred.

This study made use of deidentified data from the IC/ES Data Repository, which is managed by the Institute for Clinical Evaluative Sciences with support from its funders and partners: Canada’s Strategy for Patient-Oriented Research (SPOR), the Ontario SPOR Support Unit, the Canadian Institutes of Health Research, and the Government of Ontario. The opinions, results, and conclusions reported are those of the authors. No endorsement by IC/ES or any of its funders or partners is intended or should be inferred.

Parts of this material are based on data and/or information compiled and provided by CIHI and the Ontario Ministry of Health. The analyses, conclusions, opinions, and statements expressed herein are solely those of the authors and do not reflect those of the funding or data sources; no endorsement is intended or should be inferred.

The authors thank IQVIA Solutions Canada Inc. for use of their Drug Information File.

Parts of this material are based on data and information provided by Ontario Health (OH). The opinions, results, view, and conclusions reported in this article are those of the authors and do not necessarily reflect those of OH. No endorsement by OH is intended or should be inferred.

This document used data adapted from the Statistics Canada Postal CodeOM Conversion File, which is based on data licensed from Canada Post Corporation, and/or data adapted from the Ontario Ministry of Health Postal Code Conversion File, which contains data copied under license from ©Canada Post Corporation and Statistics Canada.

Data Availability Statement

Data can be provided on reasonable request.

Authors’ Contributions

A.R.-L.: Conceptualization, methodology, and writing—review and editing. S.B.H.: Conceptualization, methodology, and writing—review and editing. R.R.-L.: Conceptualization, methodology, and writing—review and editing. Y.P.: Conceptualization, methodology, writing—review and editing, supervision, and project administration.

Author Disclosure Statement

S.B.H. reports consulting with Abbott, Bayer, Dexcom, Eli Lilly, Novo Nordisk, and Sanofi; and research grants from BIO, Eli Lilly, Novo Nordisk, and Novartis. A.R.-L. reports research support from Sanofi Canada and Sanofi Global; consulting for Novo Nordisk Global, Dexcom Canada, and Sanofi United States; advisory boards for Sanofi United States; and paid presentations for Sanofi Canada, Sanofi Global, Novo Nordisk Canada, Eli Lilly Canada, Abbott Canada, Dexcom Canada, Diabetes Canada, and the American Diabetes Association. R.R.-L. reports research grants from Diabetes Canada, Astra-Zeneca, E Lilly, Cystic Fibrosis Canada, CIHR, FFRD, Janssen, JDRF, Merck, NIH, Novo Nordisk, Société Francophone du Diabète, Sanofi-Aventis, and Vertex Pharmaceutical; consulting/advisory panels for Abbott, Astra-Zeneca, Bayer, Boehringer I, Dexcom, E Lilly, HLS therapeutics, INESSS, Insulet, Janssen, Medtronic, Merck, Novo Nordisk, Pfizer, and Sanofi-Aventis; honoraria for conferences from Abbott, Astra-Zeneca, Boehringer I, CPD Network, Dexcom, CMS Canadian Medical & Surgical Knowledge Translation Research group, E Lilly, Janssen, Medtronic, Merck, Novo Nordisk, Sanofi-Aventis, Tandem, and Vertex Pharmaceutical; consumable gifts (in kind) from E Lilly and Medtronic; unrestricted grants for clinical and educational activities from Abbott, Dexcom, E Lilly, Medtronic, Merck, Novo Nordisk, and Sanofi-Aventis; patents for T2DM risk biomarkers and catheter life; and purchase fees from E Lilly (artificial pancreas). Y.P. is an employee and shareholder of Abbott.

Funding Information

This study, the preparation of the article, and publication fees were funded by Abbott.

Supplementary Material

Supplementary Table S1

Supplementary Table S2

Supplementary Table S3

References

Bommer

, Sagalova

, Heesemann

, et al. Global economic burden of diabetes in adults: Projections from 2015 to 2030. Diabetes Care, 2018; 41(5):963–970; doi: 10.2337/dc17-1962

Diabetes Canada. New data shows diabetes rates and economic burden on families continue to rise in Ontario. 2019. Available from: https://www.diabetes.ca/media-room/press-releases/new-data-shows-diabetes-rates-and-economic-burden-on-families-continue-to-rise-in-ontario [Last accessed: June 11, 2024 ].

Health Canada. Universal Access to Diabetes Medications, and Diabetes Device Fund for Devices and Supplies. 2024. Available from: https://www.canada.ca/en/health-canada/news/2024/02/universal-access-to-diabetes-medications-and-diabetes-device-fund-for-devices-and-supplies.html [Last accessed: October 5, 2024 ].

Benoit

, Hora

, Pasquel

, et al. Trends in emergency department visits and inpatient admissions for hyperglycemic crises in adults with diabetes in the U.S., 2006-2015. Diabetes Care, 2020; 43(5):1057–1064; doi: 10.2337/dc19-2449

McCoy

, Lipska

, Van Houten

, et al. Association of cumulative multimorbidity, glycemic control, and medication use with hypoglycemia-related emergency department visits and hospitalizations among adults with diabetes. JAMA Netw Open, 2020; 3(1):e1919099; doi: 10.1001/jamanetworkopen.2019.19099

ADA Standards of Care. 6. Glycemic Goals and Hypoglycemia: Standards of Care in Diabetes—2024. Diabetes Care, 2024; 47(Suppl 1):S111–S125; doi: 10.2337/dc24-S006

Guerci

, Levrat-Guillen

, Vicaut

, et al. Reduced acute diabetes events after Freestyle Libre System initiation in people 65 years or older with type 2 diabetes on intensive insulin therapy in France. Diabetes Technol Ther, 2023; 25(6):384–394; doi: 10.1089/dia.2023.0034

Guerci

, Roussel

, Levrat-Guillen

, et al. Important decrease in hospitalizations for acute diabetes events following Freestyle Libre System initiation in people with type 2 diabetes on basal insulin therapy in France. Diabetes Technol Ther, 2023; 25(1):20–30; doi: 10.1089/dia.2022.0271

Coons

, Greiver

, Aliarzadeh

, et al. Is glycemia control in Canadians with diabetes individualized? A cross-sectional observational study. BMJ Open Diabetes Res Care, 2017; 5(1):e000316; doi: 10.1136/bmjdrc-2016-000316

10.

Brazeau

, Messier

, Talbo

, et al.; BETTER Study Group. Self-reported severe and nonsevere hypoglycemia in type 1 diabetes: Population surveillance through the better patient engagement registry: Development and baseline characteristics. Can J Diabetes, 2022; 46(8):813–821; doi: 10.1016/j.jcjd.2022.05.010

11.

Rodbard

. Continuous glucose monitoring: A review of successes, challenges, and opportunities. Diabetes Technol Ther, 2016; 18(Suppl 2):S3–S13; doi: 10.1089/dia.2015.0417

12.

Carlson

, Daniel

, DeSantis

, et al. Flash glucose monitoring in type 2 diabetes managed with basal insulin in the USA: A retrospective real-world chart review study and meta-analysis. BMJ Open Diabetes Res Care, 2022; 10(1):e002590; doi: 10.1136/bmjdrc-2021-002590

13.

Evans

, Welsh

, Ells

, et al. The impact of flash glucose monitoring on glycaemic control as measured by HbA1c: A meta-analysis of clinical trials and real-world observational studies. Diabetes Ther, 2020; 11(1):83–95; doi: 10.1007/s13300-019-00720-0

14.

Haak

, Hanaire

, Ajjan

, et al. Flash glucose-sensing technology as a replacement for blood glucose monitoring for the management of insulin-treated type 2 diabetes: A multicenter, open-label randomized controlled trial. Diabetes Ther, 2017; 8(1):55–73; doi: 10.1007/s13300-016-0223-6

15.

Wada

, Onoue

, Kobayashi

, et al. Flash glucose monitoring helps achieve better glycemic control than conventional self-monitoring of blood glucose in non-insulin-treated type 2 diabetes: A randomized controlled trial. BMJ Open Diabetes Res Care, 2020; 8(1):e001115; doi: 10.1136/bmjdrc-2019-001115

16.

Bergenstal

, Kerr

MSD

, Roberts

, et al. Flash CGM is associated with reduced diabetes events and hospitalizations in insulin-treated type 2 diabetes. J Endocr Soc, 2021; 5(4):bvab013; doi: 10.1210/jendso/bvab013

17.

Blissett

, Blissett

, Levrat-Guillen

, et al. FreeStyle Libre Flash Glucose Monitoring system for people with type 1 diabetes in the UK: A budget impact analysis. BMJ Open Diabetes Res Care, 2022; 10(2); doi: 10.1136/bmjdrc-2021-002580

18.

ADA Standards of Care. 7. Diabetes technology: Standards of care in diabetes—2024. Diabetes Care, 2024; 47(Suppl 1):S126–S144; doi: 10.2337/dc24-S007

19.

Samson

, Vellanki

, Blonde

, et al. American Association of clinical endocrinology consensus statement: Comprehensive type 2 diabetes management algorithm – 2023 update. Endocr Pract, 2023; 29(5):305–340; doi: 10.1016/j.eprac.2023.02.001

20.

Cheng

AYY

, Feig

, Ho

, et al.; Diabetes Canada Clinical Practice Guidelines Steering Committee. Blood glucose monitoring in adults and children with diabetes: Update 2021. Can J Diabetes, 2021; 45(7):580–587; doi: 10.1016/j.jcjd.2021.07.003

21.

Diabetes Canada. Coverage of Glucose Monitoring Devices. 2024. Available from: https://www.diabetes.ca/getmedia/0170dc2c-0efe-4543-97b9-7b4563bb845f/Glucose_Monitoring_Devices_EN_2024.pdf [Last accessed: October 25, 2024 ].

22.

Diabetes Canada. Diabetes in Ontario. 2023. Available from: https://www.diabetes.ca/DiabetesCanadaWebsite/media/Advocacy-and-Policy/Backgrounder/2023_Backgrounder_Ontario.pdf [Last accessed: June 11, 2024 ].

23.

Diabetes Canada. Snapshot of Diabetes in Canada. 2023. Available from: https://www.diabetes.ca/DiabetesCanadaWebsite/media/Advocacy-and-Policy/Backgrounder/2023_Backgrounder_Canada_English.pdf [Last accessed: June 11, 2024 ].

24.

IC/ES. About IC/ES Data. 2024. Available from: https://www.ices.on.ca/about-ices-data/ [Last accessed: April 11, 2024 ].

25.

Houlden

, Thayalan

, Shi

, et al. Identifying Ontarians with type 2 diabetes mellitus in administrative data: A comparison of two case definitions. Diabetes Ther, 2024; 15(3):677–689; doi: 10.1007/s13300-024-01535-4

26.

Moin

, Troke

, Plumptre

, et al. Impact of the COVID-19 pandemic on diabetes care for adults with type 2 diabetes in Ontario, Canada. Can J Diabetes, 2022; 46(7):715–721; doi: 10.1016/j.jcjd.2022.04.009

27.

Hux

, Ivis

, Flintoft

, et al. Diabetes in Ontario: Determination of prevalence and incidence using a validated administrative data algorithm. Diabetes Care, 2002; 25(3):512–516; doi: 10.2337/diacare.25.3.512

28.

Madar

, Lalanne-Mistrih

, Lebbar

, et al. Cardiovascular risk factors and adherence to cardiovascular protection practice guidelines in adults with type 1 diabetes: A better registry cross-sectional analysis. Can J Diabetes, 2023; 47(6):473–481.e1; doi: 10.1016/j.jcjd.2023.04.006

29.

Choi

, Booth

, Jung

, et al. Association of diabetes with frequency and cost of hospital admissions: A retrospective cohort study. CMAJ Open, 2021; 9(2):e406–e412; doi: 10.9778/cmajo.20190213

30.

Wodchis

, Bushmeneva

, Nikitovic

, et al. Guidelines on Person-Level Costing Using Administrative Databases in Ontario. Toronto; 2013. Available from: https://tspace.library.utoronto.ca/handle/1807/87373 [Last accessed: October 30, 2024 ].

31.

Rapattoni

, Zante

, Tomas

, et al. A retrospective observational population-based study to assess the prevalence and burden of illness of type 2 diabetes with an estimated glomerular filtration rate < 90 mL/min/1.73 m(2) in Ontario. Diabetes Obesity Metabolism, 2021; 23(4):916–928; doi: 10.1111/dom.14294

32.

Roussel

, Riveline

, Vicaut

, et al. Important drop in rate of acute diabetes complications in people with type 1 or type 2 diabetes after initiation of flash glucose monitoring in France: The RELIEF Study. Diabetes Care, 2021; 44(6):1368–1376; doi: 10.2337/dc20-1690

33.

Karter

, Parker

, Moffet

, et al. Association of real-time continuous glucose monitoring with glycemic control and acute metabolic events among patients with insulin-treated diabetes. JAMA, 2021; 325(22):2273–2284; doi: 10.1001/jama.2021.6530

34.

Hannah

, Nemlekar

, Green

, et al. Reduction in diabetes-related hospitalizations and medical costs after DEXCOM G6 continuous glucose monitor initiation in people with type 2 diabetes using intensive insulin therapy. Adv Ther, 2024; 41(6):2299–2306; doi: 10.1007/s12325-024-02851-8

35.

Miller

, Roberts

, Joseph

, et al. Comparing clinical outcomes between two continuous glucose monitors: Similar diabetes-related events, all-cause hospitalizations and HbA1c reductions in type 1 and type 2 diabetes. Am J Med Open, 2023; 9:100008; doi: 10.1016/j.ajmo.2022.100008

36.

McCoy

, Lipska

, Van Houten

, et al. Paradox of glycemic management: Multimorbidity, glycemic control, and high-risk medication use among adults with diabetes. BMJ Open Diabetes Res Care, 2020; 8(1); doi: 10.1136/bmjdrc-2019-001007

37.

Ivers

, Jiang

, Alloo

, et al. Diabetes Canada 2018 clinical practice guidelines: Key messages for family physicians caring for patients living with type 2 diabetes. Can Fam Physician, 2019; 65(1):14–24.

38.

Meneilly

, Knip

, Miller

, et al.; Diabetes Canada Clinical Practice Guidelines Expert Committee. Diabetes in older people. Can J Diabetes, 2018; 42(Suppl 1):S283–S295; doi: 10.1016/j.jcjd.2017.10.021

39.

Shah

, Lipscombe

, Booth

. Glycemic control among people with diabetes in Ontario: A population-based cross-sectional study. Can J Diabetes, 2021; 45(4):313–318; doi: 10.1016/j.jcjd.2020.09.017

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.

0.00 MB

0.13 MB

0.03 MB

0.13 MB