Sage Journals: Discover world-class research

Abstract

Background:

Socioeconomic status (SES) and ethnic disparities in diabetes technology use have not been comprehensively explored across Canada. We describe SES disparities in continuous glucose monitoring (CGM) use among children in three Canadian provinces with differing public funding structures.

Methods:

We conducted a case–control study of children aged 1–18 years with type 1 diabetes using clinical data from three diabetes centers in Ontario, Alberta, and Québec. We measured SES using validated national neighborhood-level dimensions (residential instability, economic dependency [including employment], ethnocultural composition, and situational vulnerability [including education]). Cases were those with first-time CGM use in 2017–2022; controls were those without such use as of their last visit in that period. We examined the association between SES and CGM use using multilevel logistic regression with random effects for province, adjusting for age, sex, diabetes duration, insulin pump use, and average hemoglobin A1c.

Results:

We identified 1770 children, 48.9% female, with median (interquartile range) age 11.8 (8.8–14.3) years and duration of diabetes 2.6 (0.7–5.9) years. Of the 1770 children, 1411 (79.7%) used CGM. We observed significant associations with CGM use for three of the four SES dimensions. Compared with the least deprived quintiles for economic dependency, those in the middle quintile had 17% higher odds (adjusted odds ratio [aOR] 1.17, 95% confidence interval [CI] 1.02, 1.34) of using CGM. The most versus least diverse ethnocultural composition quintiles had 37% lower odds (aOR 0.63, 95% CI 0.60, 0.67) of using CGM, and those in the most versus least deprived quintiles for situational vulnerability (including least well-educated) had 50% lower odds (aOR 0.50, 95% CI 0.40, 0.62) of using CGM.

Conclusions and Relevance:

We found significant associations of employment status, ethnocultural diversity, and education with CGM use across Canada. Future work should promote equitable technology use among all groups.

Introduction

Type 1 diabetes is one of the most common chronic conditions in childhood and is associated with significant morbidity.¹ Insulin pumps and continuous glucose monitoring (CGM) are now part of standard of care for type 1 diabetes and can improve glycemic management and quality of life.^2–4 However, access to, and use of, these technologies is not equitable across socioeconomic status (SES), as has been shown in several countries with differing health care systems.^5,6 In Canada, these disparities have been less comprehensively studied.

Although all 10 Canadian provinces provide financial coverage for medically necessary physician and hospital-based services, coverage of diabetes technology varies. All Canadian provinces have implemented universal pediatric insulin pump funding programs providing varying amounts of financial assistance, but each province offers different CGM coverage, which has also evolved over time. Within this inconsistent funding landscape, SES disparities in insulin pump uptake and glycemic management have been found,^7–9 but disparities in CGM use have not been well explored beyond single-center studies, limiting the generalizability of previous research.^10–12 To address this limitation, our multicenter collaboration provided the first opportunity to study SES and ethnic disparities in CGM use across three different Canadian provinces, each with a distinct public funding structure for CGM.

During the time frame of our study, Ontario covered the full cost of intermittently scanned CGM for children with type 1 diabetes and no private health insurance as of September 2019, Québec provided coverage for real-time CGM for all children aged 2 years and older with type 1 diabetes as of May 2021, and Alberta did not provide any coverage for CGM. Given these differing funding structures, we hypothesized that neighborhood-level SES disparities would exist in CGM use in children with type 1 diabetes across Canada and that these disparities would differ by province. We further hypothesized that, in addition to disparities in financial security or employment, there would be education and ethnicity disparities in CGM use, irrespective of funding structure, as have been found in studies in other countries.^2,5,6 As such, our multicenter study provided a unique opportunity to better understand SES disparities within the context of a public health funding system and to inform targeted interventions aimed at improving health equity and access to technology for all children with type 1 diabetes.

Research Design and Methods

Study design and data sources

We conducted a case–control study using pooled, deidentified clinical data from three institutions in three provinces: Children’s Hospital of Eastern Ontario (Ontario), Alberta Children’s Hospital (Alberta), and Montreal Children’s Hospital (Québec). Data were extracted directly from the electronic medical record and included demographic, clinical, treatment, and laboratory information that was collected at and between clinic visits. All individuals followed in each institution’s diabetes clinic were eligible for inclusion. Informed consent/assent was obtained from participants at Alberta Children’s Hospital and Montreal Children’s Hospital but was not required at Children’s Hospital of Eastern Ontario as data were collected prospectively as part of routine clinical care. Our study received institutional review board approval from all three institutions.

Case and control identification

From the pooled data, we identified individuals aged 1–18 years with type 1 diabetes, as documented by their health care provider based on clinical characteristics and/or autoantibody testing if there was diagnostic uncertainty. Individuals were included in our study as of January 2017 (for those diagnosed on or prior to that date) or at their date of diagnosis if after 2017. Those using CGM prior to January 1, 2017, were excluded, as Health Canada first approved CGM devices in late 2016 and early 2017.^13,14 Those using CGM prior to 2017 may therefore have been using them in a research setting or purchased from other countries. Additional exclusion criteria were missing postal code, missing date of diagnosis or date of birth, or residence outside of the province in which medical care was received (Fig. 1).

FIG. 1.

Identification of included individuals from each province.

Cases were individuals who used CGM during the study period, including both intermittently scanned and real-time systems. The first recorded date of CGM use was considered the case index date. Controls were those who did not use CGM at any time during the study period; the last recorded visit date within the study period (September 2021 [Ontario], December 2021 [Alberta], or January 2022 [Québec]) was considered the control index date.

Primary exposure

Our primary exposure was neighborhood-level SES which was determined using the national-level validated 2016 Canadian Index of Multiple Deprivation (CIMD).¹⁵ CIMD is widely used by researchers and policy makers, and it is recommended by Statistics Canada for the study of marginalization and inequities.¹⁶ CIMD is derived directly from national census data and thus does not require additional questionnaires or burden for individuals. As these data are already collected through the census, CIMD not only allows for a broader understanding of inequities but also provides a reliable measure of deprivation when it is not possible to obtain individual measures of SES through surveys.

CIMD is divided into four dimensions of deprivation: residential instability, economic dependency, ethnocultural composition, and situational vulnerability. Residential instability includes, among other indicators, the proportion of dwellings that are owned, the proportion of persons living alone, and the proportion of persons who moved within the last 5 years. Economic dependency includes, among other indicators, the ratio of employment to the population and the dependency ratio (population aged 0–14 years or 65 years and older divided by population aged 15–64 years). Ethnocultural composition includes the proportion of persons who self-identify as being visibly minoritized, are foreign-born, are recent immigrants, and who have no knowledge of either official national language (English or French). Situational vulnerability includes the proportion of the population aged 25–64 years without a high school diploma, the proportion who identify as Indigenous, and the proportion of dwellings needing major repairs. Within a given dimension, quintile (Q) 1 represents those least deprived and Q5 those most deprived. To assign CIMD deprivation quintiles, six-digit postal codes were first linked to dissemination areas (DAs; small geographic units of 400–700 people) using the Postal Code Conversion File Plus¹⁷; DAs were then linked to deprivation quintiles using publicly available CIMD datasets.¹⁵ As has been done in other studies,^10,18 CIMD quintiles were used as categorical variables for each dimension, grouping Q1 and Q2 together, Q3 alone, and Q4 and Q5 together.

Covariables

Covariables were age, sex, duration of diabetes, insulin pump use, and average hemoglobin A1c (HbA1c), all determined at the index date. Sex and insulin pump use were included as categorical variables. Age, average HbA1c, and duration of diabetes were included as continuous variables. Average HbA1c was calculated by averaging all available HbA1c in the 12 months prior to and including the index date but excluding HbA1c within 3 months of diagnosis. Individuals diagnosed within 3 months of their index date were considered to have missing HbA1c.

Statistical analysis

We calculated descriptive statistics including proportions for categorical variables and median and interquartile range (IQR) for continuous variables given nonnormality. For comparisons between cases and controls, we performed chi-square tests for categorical variables and Wilcoxon Mann–Whitney tests for continuous variables. We conducted multilevel logistic regression models using generalized estimating equations and random effects for clusters (province) to examine the association between SES and CGM use, first as a crude analysis with CIMD without other covariables and then as an adjusted analysis including all covariables listed above. All odds ratios (OR) were reported with associated 95% confidence interval (CI). Multiple imputation by chained equations was used to assign average HbA1c for individuals with missing data using predictive mean matching; five replicates were generated.¹⁹ A sensitivity analysis was performed limiting cases and controls to those with duration of type 1 diabetes of more than 1 year (outside of the typical honeymoon period during which time blood glucose variability is often less and HbA1c is often lower). Statistical analyses were performed using SAS, version 9.4 (SAS Institute Inc) using a two-sided significance threshold of P < 0.05.

Results

Characteristics of cases and controls

We identified 1770 individuals with type 1 diabetes, 1411 (79.7%) of whom used CGM (Fig. 1; Table 1). The median (IQR) age was 11.8 (8.8–14.3) years with median duration of diabetes 2.6 (0.7–5.9) years. Approximately half of individuals were female, and 29.3% were using insulin pumps. Median HbA1c was 8.0% (7.3%–8.9%). All three provinces were evenly represented. Approximately half of individuals were in the least deprived quintiles (Q1 and Q2) for residential instability, economic dependency, and situational vulnerability, whereas just over half of individuals were in the most diverse quintiles (Q4 and Q5) for ethnocultural composition.

Table 1.

Characteristics of Cases and Controls

Characteristic	Total(N = 1770)	CGM users(N = 1411)	Non-CGM users(N = 359)	P-value^a
CGM use, N (%)	1411 (79.7)	1411 (100.0)	0 (0.0)	—
Age (years), median (IQR)	11.8 (8.8–14.3)	11.4 (8.4–13.8)	13.6 (10.7–15.9)	<0.001
Sex, N (%)
Male	904 (51.1)	700 (49.6)	204 (56.8)	0.01
Female	866 (48.9)	711 (50.4)	155 (43.2)	0.01
Duration of diabetes, median (IQR)	2.6 (0.7–5.9)	2.3 (0.6–5.4)	4.3 (1.6–7.9)	<0.001
Insulin pump use, N (%)
No	1251 (70.7)	946 (67.0)	305 (85.0)	<0.001
Yes	519 (29.3)	465 (33.0)	54 (15.0)	<0.001
Average HbA1c, median (IQR)	8.0% (7.3%–8.9%)	8.0% (7.3%–8.7%)	8.4% (7.4%–9.8%)	<0.001
Province
Québec	493 (27.9)	389 (27.6)	104 (29.0)	0.60
Ontario	602 (34.0)	488 (34.6)	114 (31.8)
Alberta	675 (38.1)	534 (37.9)	141 (39.3)
Residential instability, N (%)^b
Q1–2	892 (50.4)	731 (51.8)	161 (44.9)	0.01
Q3	316 (17.9)	257 (18.2)	59 (16.4)
Q4–5	562 (31.8)	423 (30.0)	139 (38.7)
Economic dependency, N (%)^b
Q1–2	930 (52.5)	741 (52.5)	189 (52.7)	0.55
Q3	332 (18.8)	271 (19.2)	61 (17.0)
Q4–5	508 (28.7)	399 (28.3)	109 (30.4)
Ethnocultural composition, N (%)^b
Q1–2	407 (23.0)	344 (24.4)	63 (17.6)	<0.001
Q3	380 (21.5)	323 (22.9)	57 (15.9)
Q4–5	983 (55.5)	744 (52.7)	239 (66.6)
Situational vulnerability, N (%)^b
Q1–2	1027 (58.0)	861 (61.0)	166 (46.2)	<0.001
Q3	320 (18.1)	258 (18.3)	62 (17.3)
Q4–5	423 (23.9)	292 (20.7)	131 (36.5)

Comparisons between CGM users and nonusers, using Wilcoxon Mann–Whitney tests for continuous variables and chi-square tests for categorical variables.

For given Canadian Index of Multiple Deprivation dimension, quintile (Q) 1 is least deprived and Q5 is most deprived.

CGM, continuous glucose monitoring; IQR, interquartile range; N, number.

CGM users (cases) were younger than nonusers (controls) (11.4 [8.4–13.8] years vs. 13.6 [10.7–15.9] years, P < 0.001) and had a shorter duration of diabetes (2.3 [0.6–5.4] years vs. 4.3 [1.6–7.9] years, P < 0.001) (Table 1). Sex distribution and insulin pump use were significantly different between CGM users and nonusers (49.6% vs. 56.8% male, P = 0.01; 33.0% vs 15.0% insulin pump use, P < 0.001). Median HbA1c was lower in CGM users compared with nonusers (8.0% [7.3%–8.9%] vs. 8.4% [7.4%–9.8%], P < 0.001). Provinces were equivalently represented in both cases and controls. The distribution of quintiles for economic dependency was the same between CGM users and nonusers, but as compared with CGM users, nonusers were more concentrated in the most deprived quintiles (Q4 and Q5) for residential instability and situational vulnerability as well as in the most diverse quintiles (Q4 and Q5) for ethnocultural composition.

Individuals using CGM prior to January 1, 2017, who were excluded from the study, had a higher proportion of insulin pump use, were younger, and had lower HbA1c compared with CGM users included in this study (Supplementary Table S1).

Association between SES and CGM use

In the crude analysis, residential instability and economic dependency were not significantly associated with CGM use (Table 2). However, those in the most ethnoculturally diverse quintiles (Q4 and Q5) had 43% lower odds (crude OR 0.57, 95% CI 0.56, 0.57) of using CGM than those in the least diverse quintiles (Q1 and Q2). Additionally, those in the most situationally vulnerable quintiles (Q4 and Q5) had 58% lower odds (OR 0.42, 95% CI 0.33, 0.55) of using CGM than those in the least deprived quintiles (Q1 and Q2).

Table 2.

Association of SES with CGM Use

Variable	Crude OR(95% CI)	Adjusted OR^a(95% CI)
Age (years)	—	0.88 (0.83, 0.93)
Sex
Male	—	Reference
Female	—	1.32 (1.21, 1.44)
Duration of diabetes	—	0.91 (0.85, 0.97)
Insulin pump use
No	—	Reference
Yes	—	3.41 (2.64, 4.39)
Average HbA1c	—	0.93 (0.88, 0.99)
Residential Instability^b
Q1–2	Reference	Reference
Q3	1.07 (0.75, 1.52)	1.02 (0.73, 1.42)
Q4–5	0.98 (0.62, 1.54)	0.92 (0.63, 1.34)
Economic dependency^b
Q1–2	Reference	Reference
Q3	1.32 (1.08, 1.63)	1.17 (1.02, 1.34)
Q4–5	1.06 (0.92, 1.23)	1.07 (0.91, 1.26)
Ethnocultural composition^b
Q1–2	Reference	Reference
Q3	0.94 (0.76, 1.15)	0.91 (0.73, 1.13)
Q4–5	0.57 (0.56, 0.57)	0.63 (0.60, 0.67)
Situational vulnerabilityb
Q1–2	Reference	Reference
Q3	0.78 (0.57, 1.05)	0.86 (0.66, 1.13)
Q4–5	0.42 (0.33, 0.55)	0.50 (0.40, 0.62)

Statistically significant results are in bold.

OR for CGM use. Multilevel logistic regression using generalized estimating equations and random effects for cluster (province), adjusted for age, sex, duration of diabetes, insulin pump use, average HbA1c, and all Canadian Index of Multiple Deprivation dimensions.

For given Canadian Index of Multiple Deprivation dimension, quintile (Q). Q1 is least deprived and Q5 is most deprived.

CI, confidence interval; OR, odds ratio; SES, socioeconomic status.

These findings were minimally changed in the adjusted analysis, accounting for the effects of age, sex, duration of diabetes, insulin pump use, and average HbA1c (Table 2). Residential instability was not significantly associated with CGM use. Those in the middle quintile for economic dependency (Q3) had 17% higher odds (adjusted OR [aOR] 1.17, 95% CI 1.02, 1.34) of using CGM than those in the least deprived quintiles (Q1 and Q2). However, there was no significant association of CGM use between those in the most economically deprived quintiles (Q4 and Q5) and those in the least. Ethnocultural composition (i.e., ethnicity) and situational vulnerability (including education level) remained significantly associated. Those in the most ethnoculturally diverse quintiles had 37% lower odds (aOR 0.63, 95% CI 0.60, 0.67) of using CGM than those in the least diverse quintiles. Those in the most situationally vulnerable quintiles had 50% lower odds (aOR 0.50, 95% CI 0.40, 0.62) of using CGM than those in the least deprived quintiles. Age, duration of diabetes, and average HbA1c were also significantly associated, with each increasing year of age or diabetes duration associated with 12% and 9% lower odds, respectively, of CGM use and with each increasing percentage point of average HbA1c associated with 7% lower odds of CGM use. Insulin pump use was significantly associated with increased CGM use, as was female sex.

A sensitivity analysis was performed limiting the study sample to cases and controls with duration of diabetes longer than one year. This group consisted of 1241 individuals, 948 cases and 293 controls. Of those individuals excluded, 153 had a duration of diabetes less than 90 days at their index date (142 cases and 11 controls). As in the analysis of the full data, economic dependency (Q3 only), ethnocultural composition (Q4-5 only) and situational vulnerability (Q4-5 only) remained significantly associated with CGM use whereas residential instability remained nonsignificant. Age, sex, and insulin pump use remained significant, but duration of diabetes and average HbA1c were no longer significantly associated with CGM use (Table 3).

Table 3.

Association of SES with CGM Use, Limited to Cases and Controls with Duration of Diabetes Longer Than 1 Year (N = 1241)

Variable	Crude OR(95% CI)	Adjusted OR^a(95% CI)
Age (years)	—	0.88 (0.85, 0.92)
Sex
Male	—	Reference
Female	—	1.26 (1.12, 1.41)
Duration of diabetes	—	0.93 (0.87, 1.00)
Insulin pump use
No	—	Reference
Yes	—	3.77 (2.76, 5.15)
Average HbA1c	—	0.96 (0.86, 1.07)
Residential instability^b
Q1–2	Reference	Reference
Q3	1.06 (0.64, 1.75)	1.02 (0.60, 1.73)
Q4–5	0.91 (0.52, 1.56)	0.88 (0.54, 1.43)
Economic dependency^b
Q1–2	Reference	Reference
Q3	1.32 (1.01, 1.74)	1.25 (1.04, 1.51)
Q4–5	1.14 (0.92, 1.41)	1.19 (0.94, 1.51)
Ethnocultural composition^b
Q1–2	Reference	Reference
Q3	0.81 (0.59, 1.11)	0.82 (0.58, 1.16)
Q4–5	0.55 (0.52, 0.58)	0.64 (0.61, 0.66)
Situational vulnerability^b
Q1–2	Reference	Reference
Q3	0.82 (0.47, 1.44)	0.94 (0.51, 1.76)
Q4–5	0.42 (0.30, 0.61)	0.52 (0.35, 0.77)

Statistically significant results are in bold.

Discussion

Our research on the association between SES and CGM use represents the most comprehensive study across Canada to date with inclusion of three provinces, strengthening the generalizability of our findings. In our multicenter case–control study, the majority of individuals (79.7%) were using CGM. In the setting of differing provincial financial coverage models for CGM during our study time frame, with one province (i.e., Québec) providing full coverage and another (i.e., Alberta) providing none, we found persistent associations of CGM use across all provinces with economic dependency (including employment), ethnocultural composition (i.e., ethnicity), and situational vulnerability (including education level).

We found a significant association of neighborhood-level economic dependency with CGM use, in that those of the middle quintile had higher odds of CGM use than those least deprived, but there was no significant association for those most deprived (including highest unemployment ratio). The reasons for higher use in the middle quintile in our study are not immediately evident. Our results show less pronounced disparities in employment status than prior single-center studies in Québec and Ontario, which demonstrated lower CGM use in most deprived quintiles compared with least, using a validated combined material and social deprivation index (Québec) and another validated material deprivation dimension encompassing education, employment, and income (Ontario).^10,11 A single province study from British Columbia also using CIMD showed lower CGM use in the more deprived quintile for economic dependency compared with least.¹² Similarly, in the international context, in a large pediatric registry study from the Type 1 Diabetes Exchange in the United States and the Diabetes Prospective Follow-up in Germany, CGM use was significantly higher in those least deprived compared to those most deprived, defined by either a composite score including education, insurance, and income (United States) or a validated, neighborhood-level index encompassing income, employment, education, and social capital (Germany).⁵ However, these prior studies used metrics that combined markers of financial stability with those of other SES variables (e.g., combining income, employment, and education level into one score), in contrast to the validated CIMD we used for SES in our study, which provided more discrete and detailed dimensions. Thus, our multiprovincial study findings suggest that the drivers of CGM use go beyond monetary concerns and employment status.²⁰

Our study suggests that factors such as ethnocultural composition and situational vulnerability contribute to some of the disparities observed in CGM use, with those in the most diverse quintiles (minoritized populations) and the most situationally vulnerable (including higher proportion without a high school diploma) having lower odds of CGM use than their counterparts in the least diverse and least situationally vulnerable quintiles, respectively. The role of ethnicity in CGM use in Canada has previously not been well studied and was excluded from full analyses in a recent study from British Columbia.¹² Our findings across these three Canadian provinces corroborate studies from other countries such as the United States, Germany, and New Zealand, which have shown SES disparities in both CGM and insulin pump use by ethnicity and education level, including the previously mentioned study through the Type 1 Diabetes Exchange.^5,6,21–23 A large epidemiological study of over 48,000 children and adults with type 1 diabetes in the United States demonstrated that while CGM use has increased over time, ethnic disparities have persisted with a lower percentage of those identifying as non-Hispanic Black or Hispanic using CGM than those identifying as non-Hispanic White.²⁴ In New Zealand, another country with public funding, those of non-European ethnicity had significantly lower odds of using an insulin pump compared to those of European ethnicity, although CGM use was not studied.⁶ In terms of education level, a multicenter study from the United States showed that having a parent with college education or higher was significantly associated with greater CGM use in adolescents.²⁵ Taken together, the above findings have important implications in demonstrating factors other than financial concerns contributing to the persistent disparities in technology use even in the setting of government financial coverage.

The underlying drivers for disparities in CGM use by neighborhood-level ethnocultural composition and situational vulnerability (including education level) in our study could be related to patient and family preferences as well as health care provider biases and structural racism.²⁶ Patients and families may choose not to start CGM for a variety of reasons, including reluctance to wear visible devices^27,28 and parental concerns with complexity of the technology and overwhelming amounts of data.²⁸ Numeric literacy may also play a role in the decision to use CGM.²⁹ In a survey of parents of young children with type 1 diabetes, parental diabetes-related numeracy was inversely correlated with their child’s glycemic management,³⁰ and although not assessed in that study, we hypothesize that those with lower numeric literacy may be less inclined to use CGM given the large amount of numeric data generated.

Additionally, providers may not offer CGM to families or patients they suspect are already struggling with diabetes management (which may explain our finding of lower odds of CGM use in those with higher average HbA1c) or to already marginalized populations due to unconscious biases.^31–33 For example, a survey of 39 pediatric diabetes providers in the United States showed implicit bias against public insurance in 85% of respondents when asked about prescribing CGM in hypothetical scenarios.³¹ Given that other studies, including single center investigations from Québec, Ontario, and British Columbia, have shown that CGM use can mediate SES disparities in glycemic management among children with type 1 diabetes,^10–12 future work should include qualitative studies to investigate other patient or physician barriers to use of CGM in order to understand and address biases and to increase CGM access and use for all children with type 1 diabetes.

Quality improvement (QI) initiatives have been used to decrease SES disparities in CGM use. At a single U.S. institution, a targeted QI initiative created a sustained decrease in the disparity in CGM use between different ethnicities.³⁴ A recent multicenter initiative from the Type 1 Diabetes Exchange QI Collaborative established that targeted interventions including equity/unconscious bias training for providers helped to reduce gaps in CGM use by ethnicity.³⁵ Addressing systemic issues with regards to adapting provider support for different family contexts and educational means should also be explored in the future.

As in other studies, we did find that increasing age was significantly associated with lower odds of CGM use.^21,36 Increasing duration of diabetes was also significantly associated with lower odds of CGM use in the overall analysis and trended toward such in the study sample limited to those outside of the honeymoon period, suggesting possible therapeutic inertia of adoption of new technology in those with longer standing diabetes. In contrast to other Canadian studies on technology use in children with type 1 diabetes, we found a significant association of CGM use with sex, with females having increased odds of CGM use.^7,11 However, similar to our findings, two large survey-based studies of adults in the United States have also found that females are more likely to use wearable devices related to cardiovascular health than males, the reason for which is unknown.^37,38 Although not a focus of this study, our findings suggest the need for future investigations into the underlying drivers of these observed sex disparities in diabetes technology use in children and adolescents.

While we conducted a large study comprising three different centers in three distinct provinces, our study does have limitations. Clinical data were extracted directly from the electronic medical record, and as such, available data were dependent on documentation by health care providers. However, these data have the benefit of providing individual diagnostic information and longitudinal laboratory values, which may be unavailable in other data sources such as health administrative data.⁷ Although the majority of individuals in this study were using CGM, the group of nonusers was still large enough to sufficiently represent all included variables (as seen in Table 1). Our study was also limited to a single tertiary care center within each province, but the majority of youth in Canada with type 1 diabetes receive their diabetes care in these specialized centers. Moreover, SES was defined at a neighborhood rather than individual level, which may risk exposure misclassification. A previous study from the United States showed that neighborhood-level SES was better correlated in urban areas than rural areas with individual-level SES (in terms of self-reported highest education level and relative income).³⁹ Additionally, although attributing individual ethnicity based on this ecological construct is potentially more problematic than attributions based on other CIMD dimensions, the ethnocultural composition dimension used in this study has been used previously in numerous other studies for similar purposes.^8,9,40 A strength of our study, compared with other Canadian studies, is that we measured ethnicity.^11,18,40 The situational vulnerability dimension was designed with seemingly disparate measures (e.g., proportion of a certain age without a high school diploma and proportion who identify as Indigenous), which made it difficult to tease out individual effects, although we suspect many factors play a role in the observed disparities. Future research should be conducted using self-reported measures of ethnicity and highest educational level attained, as well as using other social determinants of health such as food or childcare insecurity,⁴¹ to provide a more comprehensive picture of factors affecting technology use in children with type 1 diabetes across Canada. Moreover, since our study was completed, the funding landscape for CGM has become more uniform among Ontario, Québec, and Alberta, with all three provinces offering coverage for real-time CGM for pediatric patients. However, this uniformity would not be expected to change our results as we still found neighborhood-level disparities even accounting for the random effects of province with the prior variability in coverage of CGM. Moreover, a recent study from another Canadian province (British Columbia) that also utilized CIMD as a measure of SES showed that neighborhood-level SES disparities persisted even after implementation of public health coverage,¹² thus suggesting that our study findings are generalizable across Canada.

Conclusions

While the majority of children with type 1 diabetes in three Canadian centers were using CGM, our results have several implications. First, we demonstrated some disparities by employment status even in the setting of government financial coverage for CGM, suggesting that full insurance coverage may be necessary, but is not sufficient, to minimize inequities in technology use. Second, we found neighborhood-level disparities in CGM use in terms of ethnocultural composition and situational vulnerability across Canada, suggesting that future studies also need to address explicit and implicit biases, structural barriers, and education inequities. Especially with increasing use of automated insulin delivery systems which require CGM use, research and implemented policies at a local and national level to decrease barriers to use of this standard of care technology in minoritized and less well-educated populations can help to optimize clinical care and outcomes for all children with type 1 diabetes.

Footnotes

Acknowledgments

We thank the patients and families for their participation in our clinical registries. We also appreciate Dr. Patricia Li’s feedback on earlier versions of this article.

Authors’ Contributions

J.M.L. and M.N. conceptualized the study, interpreted the data, and wrote the first draft of the article. E.R. and M.D. provided statistical analysis, data interpretation, and reviewed and edited the article. C.Z., E.B.G., R.S., D.P., and J.V.O. reviewed and provided substantive edits to earlier versions of the article. All authors approved the final version of the article. M.N. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Author Disclosure Statement

R.S. has received speaking and advisory board fees from Dexcom Canada. C.Z. has received advisory board fees from Dexcom Canada and a grant to the institution from Tandem Diabetes Care. None of the other authors have conflicts of interest to disclose.

Funding Information

J.M.L. received prior salary support from the Canadian Pediatric Endocrine Group and the Pediatric Endocrine Society. M.N. holds a Clinician-Scientist Award from the Fonds de recherche du Québec—Santé.

Supplementary Material

Abbreviations Used

References

Wherrett

, Ho

, Huot

, et al.; Diabetes Canada Clinical Practice Guidelines Expert Committee. Type 1 diabetes in children and adolescents. Can J Diabetes, 2018; 42(Suppl 1):S234–S246; doi: 10.1016/j.jcjd.2017.10.036

Paris

, Imperatore

, Klingensmith

, et al. Predictors of insulin regimens and impact on outcomes in youth with type 1 diabetes: The SEARCH for Diabetes in Youth study. J Pediatr, 2009; 155(2):183–189.e1; doi: 10.1016/j.jpeds.2009.01.063

Misso

, Egberts

, Page

, et al. Continuous subcutaneous insulin infusion (CSII) versus multiple insulin injections for type 1 diabetes mellitus. Cochrane Database Syst Rev, 2010; 1(1):Cd005103; doi: 10.1002/14651858.CD005103.pub2

Laffel

, Kanapka

, Beck

, et al.; CDE10. Effect of continuous glucose monitoring on glycemic control in adolescents and young adults with type 1 diabetes: A randomized clinical trial. JAMA, 2020; 323(23):2388–2396; doi: 10.1001/jama.2020.6940

Addala

, Auzanneau

, Miller

, et al. A decade of disparities in diabetes technology use and HbA1c in pediatric type 1 diabetes: A transatlantic comparison. Diabetes Care, 2021; 44(1):133–140; doi: 10.2337/dc20-0257

Wheeler

, Braund

, Galland

, et al. District health board of residence, ethnicity and socioeconomic status all impact publicly funded insulin pump uptake in New Zealand patients with type 1 diabetes. The New Zealand Medical J, 2019; 132(1491):78–89.

Ladd

, Sharma

, Rahme

, et al. Comparison of socioeconomic disparities in pump uptake among children with type 1 diabetes in 2 Canadian Provinces with different payment Models. JAMA Netw Open, 2022; 5(5):e2210464; doi: 10.1001/jamanetworkopen.2022.10464

Shulman

, Stukel

, Miller

, et al. Insulin pump use and discontinuation in children and teens: A population-based cohort study in Ontario, Canada. Pediatr Diabetes, 2017; 18(1):33–44; doi: 10.1111/pedi.12353

Zuijdwijk

, Cuerden

, Mahmud

. Social determinants of health on glycemic control in pediatric type 1 diabetes. J Pediatr, 2013; 162(4):730–735; doi: https://dx.doi.org/10.1016/j.jpeds.2012.12.010

10.

Simba

, Von Oettingen

, Rahme

, et al. Socioeconomic disparities in glycemic control in children and youth with type 1 diabetes: A retrospective cohort study. Can J Diabetes, 2023; 47(8):658–664.e2; doi: 10.1016/j.jcjd.2023.07.005

11.

Stanley

, Clarke

ABM

, Shulman

, et al. Mediating effects of technology-based therapy on the relationship between socioeconomic status and glycemic management in pediatric type 1 diabetes. Diabetes Technol Ther, 2023; 25(3):186–193; doi: 10.1089/dia.2022.0388

12.

Bone

, Leach

, Addala

, et al. The impact of public policy on equitable access to technology for children and youth living with type 1 diabetes in British Columbia, Canada. Diabetes Technol Ther, 2025; 27(3):194–201; doi: 10.1089/dia.2024.0366

13.

Dexcom. With Health Canada approval, Dexcom G5® Mobile CGM System is the First Medical Device in North America for Making Daily Diabetes Decisions Without Painful Fingersticks. 2016. Available from: https://www.dexcom.com/news/health-canada-dexcom-g5-mobile-approval [Last accessed: August 6, 2023].

14.

CADTH. Informing decisions about new health technologies: Flash glucose monitoring system for diabetes. 2018. Available from: https://www.cadth.ca/sites/default/files/pdf/eh0053_flash_glucose_monitoring_system_for_diabetes_mar2018.pdf

15.

Statistics Canada. The Canadian Index of Multiple Deprivation. Statistics Canada Catalogue, 2019 no. 45-20-0001

16.

Statistics Canada. The Canadian Index of Multiple Deprivation User Guide. 2021. Available from: https://www150.statcan.gc.ca/n1/pub/45-20-0001/452000012023002-eng.htm [Last accessed: March 18, 2025].

17.

Statistics Canada. Postal Code Conversion File Plus. 2017. Available from: https://www150.statcan.gc.ca/n1/en/catalogue/82F0086X2017001 [Last accessed: August 6, 2023].

18.

Nakhla

, Rahme

, Simard

, et al. Risk of ketoacidosis in children at the time of diabetes mellitus diagnosis by primary caregiver status: A population-based retrospective cohort study. CMAJ, 2018; 190(14):E416–E421; doi: 10.1503/cmaj.170676

19.

Rubin

. Multiple imputation after 18+ years. J American Statistical Association, 1996; 91(434):473–489; doi: 10.1080/01621459.1996.10476908

20.

Statistics Canada. The Canadian Index of Multiple Deprivation Dataset. Available from: https://www150.statcan.gc.ca/n1/pub/45-20-0001/452000012019001-eng.htm

21.

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; doi: 10.1089/dia.2018.0384

22.

Monkemoller

, Muller-Godeffroy

, Lilienthal

, et al. The association between socio-economic status and diabetes care and outcome in children with diabetes type 1 in Germany: The DIAS study (diabetes and social disparities). Pediatr Diabetes, 2019; 20(5):637–644; doi: 10.1111/pedi.12847

23.

Lai

, Lipman

, Willi

, et al. Racial and ethnic disparities in rates of continuous glucose monitor initiation and continued use in children with type 1 diabetes. Diabetes Care, 2021; 44(1):255–257; doi: 10.2337/dc20-1663

24.

Ebekozien

, Mungmode

, Sanchez

, et al. Longitudinal trends in glycemic outcomes and technology use for over 48,000 people with type 1 diabetes (2016-2022) from the T1D exchange quality improvement collaborative. Diabetes Technol Ther, 2023; 25(11):765–773; doi: 10.1089/dia.2023.0320

25.

Chen

, Tinsley

, Volkening

, et al. Observed characteristics associated with diabetes device use among teens with type 1 diabetes. J Diabetes Sci Technol, 2023; 17(1):186–194; doi: 10.1177/19322968211050069

26.

Hawkes

, Lipman

. Racial disparities in pediatric type 1 diabetes: Yet another consequence of structural racism. Pediatrics, 2021; 148(2); doi: 10.1542/peds.2021-050333

27.

Commissariat

, Whitehouse

, Hilliard

, et al. Sources and valence of information impacting parents’ decisions to use diabetes technologies in young children <8 years old with type 1 diabetes. Diabetes Technology & Therapeutics, 2020; 22(9):697–700; doi: 10.1089/dia.2019.0497

28.

Hilliard

, Levy

, Anderson

, et al. Benefits and barriers of continuous glucose monitoring in young children with type 1 diabetes. Diabetes Technol Ther, 2019; 21(9):493–498; doi: 10.1089/dia.2019.0142

29.

Anhalt

. Limitations of continuous glucose monitor usage. Diabetes Technol Ther, 2016; 18(3):115–117; doi: 10.1089/dia.2016.0011

30.

Pulgaron

, Sanders

, Patino-Fernandez

, et al. Glycemic control in young children with diabetes: The role of parental health literacy. Patient Educ Couns, 2014; 94(1):67–70; doi: 10.1016/j.pec.2013.09.002

31.

Addala

, Hanes

, Naranjo

, et al. Provider implicit bias impacts pediatric type 1 diabetes technology recommendations in the United States: Findings from the gatekeeper study. J Diabetes Sci Technol, 2021; 15(5):1027–1033; doi: 10.1177/19322968211006476

32.

Raphael

, Oyeku

. Implicit bias in pediatrics: An emerging focus in health equity research. Pediatrics, 2020; 145(5); doi: 10.1542/peds.2020-0512

33.

Hall

, Chapman

, Lee

, et al. Implicit racial/ethnic bias among health care professionals and its influence on health care outcomes: A systematic review. Am J Public Health, 2015; 105(12):e60-76–e76; doi: 10.2105/ajph.2015.302903

34.

Schmitt

, Fogle

, Scott

, et al. Improving equitable access to continuous glucose monitors for Alabama’s children with type 1 diabetes: A quality improvement project. Diabetes Technol Ther, 2022; 24(7):481–491; doi: 10.1089/dia.2021.0511

35.

Odugbesan

, Mungmode

, Rioles

, et al.; T1D Exchange Quality Improvement Collaborative. Increasing continuous glucose monitoring use for non-Hispanic black and Hispanic people with type 1 diabetes: Results from the T1D exchange quality improvement collaborative equity study. Clin Diabetes, 2024; 42(1):40–48; doi: 10.2337/cd23-0050

36.

Sawyer

, Sobczak

, Forlenza

, et al. Glycemic control in relation to technology use in a single-center cohort of children with type 1 diabetes. Diabetes Technol Ther, 2022; 24(6):409–415; doi: 10.1089/dia.2021.0471

37.

Chandrasekaran

, Sharma

, Moustakas

. Exploring disparities in healthcare wearable use among cardiovascular patients: Findings from a national survey. Rev Cardiovasc Med, 2023; 24(11):307; doi: 10.31083/j.rcm2411307

38.

Venn

, Khurshid

, Grayson

, et al. Characteristics and attitudes of wearable device users and nonusers in a large health care system. J Am Heart Assoc, 2024; 13(1):e032126; doi: 10.1161/jaha.123.032126

39.

Xie

, Hubbard

, Himes

. Neighborhood-level measures of socioeconomic status are more correlated with individual-level measures in urban areas compared with less urban areas. Ann Epidemiol, 2020; 43:37–43.e4; doi: 10.1016/j.annepidem.2020.01.012

40.

Shulman

, Stukel

, Miller

, et al. Low socioeconomic status is associated with adverse events in children and teens on insulin pumps under a universal access program: A population-based cohort study. BMJ Open Diabetes Res Care, 2016; 4(1):e000239; doi: 10.1136/bmjdrc-2016-000239

41.

Hershey

, Morone

, Lipman

, et al. Social determinants of health, goals and outcomes in high-risk children with type 1 diabetes. Can J Diabetes, 2021; 45(5):444–450 e1; doi: 10.1016/j.jcjd.2021.02.005

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

0.02 MB

Disparities in Continuous Glucose Monitoring Use in Children with Type 1 Diabetes Across Canada

Abstract

Background:

Methods:

Results:

Conclusions and Relevance:

Introduction

Research Design and Methods

Study design and data sources

Case and control identification

Primary exposure

Covariables

Statistical analysis

Results

Characteristics of cases and controls

Association between SES and CGM use

Discussion

Conclusions

Footnotes

Acknowledgments

Authors’ Contributions

Author Disclosure Statement

Funding Information

Supplementary Material

Abbreviations Used

References

Supplementary Material