The Impact of Public Policy on Equitable Access to Technology for Children and Youth Living with Type 1 Diabetes in British Columbia,Canada

Abstract

Objective:

Structural inequities impede technology uptake in marginalized populations living with type 1 diabetes (T1D). Our objective was to describe hemoglobin A1c (HbA_1c), time in range (TIR), and pump use to evaluate the impact of a universal funding policy for continuous glucose monitoring (CGM) across levels of deprivation in children with T1D in the Canadian province of British Columbia (BC).

Methods:

Patients with T1D and at least one outpatient visit after June 10, 2020 (1-year before universal CGM funding) who were enrolled in the BC Pediatric Diabetes Registry were included (n = 477). The Canadian Index of Multiple Deprivation (quintile 1 = least deprived; quintile 5 = most deprived) was determined using postal code. Mixed effects models were used to describe HbA_1c, TIR, and pump use, and an interrupted time series generalized additive model estimated the change in CGM use pre- and postintroduction of universal coverage.

Results:

No differences were observed among the five levels of deprivation for HbA_1c and TIR; however, for residential instability, those with the highest level of deprivation had a lower probability of pump use (−18.9%, 95% confidence interval [CI] = −26.1% to −11.7% for quintile 5 vs. 1). There was an increase in CGM uptake across all levels of deprivation 1-year after introduction of universal CGM funding. For example, the difference in sensor use from the most to least deprived situational group was −21.0% (−35.4%, −6.6%) at the time of universal coverage and shrank to −4.6% (−21.6%, 12.4%) after 12 months of coverage. However, an equity gap in CGM use persisted between the least and most deprived groups (−21.9, 95% CI = −34.5 to −9.4 for quintile 5 vs. 1 in economic dependency).

Conclusions:

Universal coverage of CGM improved uptake; however, equity gaps persisted. More research is needed to explore nonfinancial barriers to diabetes technology use in marginalized populations.

Introduction

Structural inequities in pediatric type 1 diabetes (T1D) care are associated with poor diabetes outcomes.^1

–4 The use of diabetes technology such as insulin pumps and continuous glucose monitoring (CGM) has the potential to partially mediate this effect.^5
–7 Yet, a multitude of barriers spanning public policy (i.e., health insurance), societal constructs (i.e., structural racism), social determinants of health (i.e., economic stability, education, neighborhood, and built environment),^2,8,9 and even provider bias^10,11 have impeded the uptake of diabetes technology, particularly in historically marginalized communities. This has resulted in an interval worsening of diabetes outcomes for youth from the lowest socioeconomic status (SES) groups, driven specifically by diabetes technology utilization.¹

Demonstrated by randomized controlled trials¹² and more recently using real-world data from population-based diabetes registries,^13,14 CGM has indisputably improved diabetes outcomes such as glycemic control [as measured by hemoglobin A1c (HbA_1c)] in both children and adults.^13,15 Furthermore, CGM is associated with better quality of life in children with T1D and their parents/caregivers, independent of insulin delivery technology and glycemic control.¹⁶ All people living with T1D therefore should have access to CGM, yet due to social determinants of health, too many are left behind.²

Public policies supporting universal coverage of diabetes technology have been implemented to alleviate the financial barriers to accessing diabetes technology.^13,17 While this approach has the potential to move us closer to equality—providing the same coverage opportunities to everyone regardless of individual context or needs—it is less likely to achieve equity, which involves fairness through proportional allocation. More importantly, it may not address justice, which requires removing the structures and systems that create disparities. In Canada, clinical services (i.e., physician or diabetes nurse educator outpatient visits, emergency department visits, or hospitalizations) are accessed through a public single-payer system; however, diabetes technology is largely funded through private health plans that can be financially inaccessible to many. Various policies supporting universal coverage of diabetes technology have been implemented in almost all Canadian provinces; however, research shows that these policies only partially mitigate SES disparities. For example, in the Canadian provinces of Quebec and Manitoba, material deprivation was associated with decreased pump uptake even with established universal insulin pump coverage programs.¹⁷ It is essential therefore to evaluate universal coverage policies to ensure they achieve their intended goals without inadvertently widening existing equity gaps or creating new ones.

In the Canadian province of British Columbia (BC), a universal insulin pump program for children living with T1D was implemented in 2008 and expanded to adults with insulin-dependent diabetes in 2018. More recently, in June 2021, universal funding of the Dexcom G6 CGM was implemented for all citizens living with insulin-dependent diabetes. Using an existing population-based pediatric diabetes registry, the objective of this study was to describe differences in diabetes outcomes and the impact of universal CGM funding on CGM uptake across different levels of deprivation in children living with T1D in BC, Canada. We hypothesized that differences would exist by deprivation index in diabetes outcomes, as well as CGM uptake before and after universal coverage.

Research Design and Methods

Data source and population

The BC Pediatric Diabetes Registry (BC-PDR) has been collecting sociodemographic and health data on consented children living with diabetes who received care from BC Children’s Hospital (BCCH) Diabetes Clinic (consent rate ∼80%).¹⁸

BCCH Diabetes Clinic cares for roughly half of all children living with diabetes in BC, with the other half receiving care in community-based diabetes clinics. The BC-PDR collects patient- and visit-level data prospectively from the time of enrollment until the transition to adult care. Patient-level data, collected at the time of patient enrollment, include assigned sex, month and year of birth, date of diabetes diagnosis, type of diabetes, self-reported ethnicity, chronic comorbidities, presentation at onset (i.e., diabetic ketoacidosis, hyperglycemia and/or ketosis, screening test for diabetes), and HbA_1c at diagnosis. Visit-level data are collected at each outpatient clinic visit and include height, weight, body mass index, blood pressure, point of care or laboratory HbA_1c, frequency of self-monitoring of blood glucose, technology use (sensor, insulin pump), and CGM metrics, including time in range (TIR). Postal code is collected and maintained in a separate, password-protected crosswalk file and used to calculate population-based geographically based variables such as the deprivation indices.

Inclusion criteria

We included all patients with at least one outpatient visit after June 10, 2020, which is 1 year before the launch of universal funding of CGM in BC. Data were collected up until December 7, 2022, at the time of analysis. Patients with a diagnosis of non-T1D or whose postal code was not available in the BC-PDR for categorization of deprivation were excluded.

Canadian Index of Multiple Deprivation

The Canadian Index of Multiple Deprivation (CIMD) allows for assessment of inequality through multiple sources. There are four dimensions to the index as follows: economic dependency, residential instability, ethnocultural composition, and situational vulnerability (Supplementary Table S1). Data are provided from Statistics Canada at the dissemination area level, a geographically stable unit composed of 400–700 people. Indicators are derived from census data, and scores for each dimension are derived via factor analysis.¹⁹ Subsequently, quintile rankings within each dimension are created when factor scores are ordered from smallest to largest and then divided into five equally sized groups and categorized from 1 (least deprived) through 5 (most deprived).

We used postal code data from the BC-PDR to link dissemination area-level measures of deprivation to each individual. CIMD data were derived from the 2016 census, and we used the most recent postal code files from Statistics Canada, which are from 2020.

Exposures, outcomes, and covariates

First, we assessed the relationship between CIMD and several visit-level clinical outcomes, including HbA_1c (%), TIR (%), pump use (%), number of self-reported episodes of diabetic ketoacidosis (DKA) after diagnosis of T1D, and number of self-reported episodes of severe hypoglycemia (defined as hypoglycemia resulting in seizure or loss of consciousness or requiring assistance for management). We did not include results for the CIMD dimension depicting ethnocultural composition as we had concerns about its validity in the province of BC where immigrant populations and visible minorities are likely distributed across many dissemination areas. We adjusted the analysis for age at diagnosis, age at measurement, and patient-assigned sex.

Second, we exploited the province-wide implementation of universal coverage for CGM to conduct an interrupted time series analysis.²⁰ For this objective, we compared the trends in CGM use (assessed at each visit) preuniversal coverage to those postuniversal coverage, stratified by CIMD quintiles, to assess whether those with higher (compared with lower) levels of deprivation had differential benefit from universal coverage.

Statistical analysis

The included cohort was described using medians and interquartile ranges for continuous variables and counts and percentages for categorical variables. We further stratified data by quintile and provided descriptive data on variables used in regression analyses (see below).

HbA_1c, TIR, and pump use were summarized by CIMD dimension and quintile graphically. As patients may contribute more than one measure during the study period, we estimated possible differences in these outcomes between CIMD quintiles using generalized linear mixed effects models. For HbA_1c and TIR, we used linear mixed effects models, whereas for pump use, we used a logistic generalized additive model (GAM) with random effects for subject. Results for HbA_1c and TIR are expressed as mean differences between CIMD levels, whereas differences in pump use were summarized using risk differences derived from the logistic model.²¹ Models were adjusted for the variables referred to above. All estimates above are presented with corresponding 95% confidence intervals (CIs). Due to small sample sizes, only descriptive analyses for rates of DKA and severe hypoglycemia by CIMD quintile were conducted.

To assess the change in CGM use pre- and postintroduction of universal CGM coverage, we used a segmented logistic GAM. The advantage of a GAM is that it naturally allows for possibly nonlinear trends in CGM use over time.²² The date of implementation of universal coverage was considered the “interruption,” and the model was segmented at this time point (i.e., representing a possible change in trend). To allow for differences by deprivation level, an interaction between deprivation quintile and both pre- and post-CGM coverage trend was included. Results are summarized as fitted values from this model graphically by quintile. We further calculated the adjusted marginal risk difference in sensor use between each deprivation quintile and the highest: (1) in the overall study period, (2) at the time of implementation of universal coverage, and (3) 12 months following coverage implementation. The delta method was used to estimate 95% CIs for the risk differences.

All analyses were conducted using R statistical software version 4.3.0.²³

Results

At the time of analysis, the BC-PDR contained 681 patients. Of these, a postal code was not available in 48 (7.0%) patients. Of the remaining 633 patients, 584 (92.3%) had a diagnosis of T1D, of whom 477 (81.6%) had at least one outpatient visit after June 10, 2020.

The cohort of 477 patients is described in Table 1. Briefly, they were slightly more likely to have male assigned sex (57.9%), with a median age of diagnosis of 6.6 years and a median age at study start of 13.2 years. Patients had a median of 5 visits throughout the study period. The largest proportion of patients in the economic instability and situational vulnerability dimensions (31.9% and 34.8% of patients, respectively) were in the least deprived quintile (first), whereas for the residential instability dimension, the middle quintiles (second, third, and fourth) accounted for the greatest proportion of patients. Demographics were similar across quintiles (Supplementary Table S2).

Table 1.

Demographics, Deprivation Quintiles, and Visit Frequency in Study Population

	Total (N = 477)
Sex, n (%)
Female	201 (42.1)
Male	276 (57.9)
Age at diagnosis (years), median [IQR]	6.6 [3.7, 10.1]
Age at study start (years), median [IQR]	13.2 [9.7, 16.2]
Ethnicity, n (%)
White	255 (62.3%)
South Asian	33 (8.1%)
Southeast Asian	11 (2.7%)
East Asian	14 (3.4%)
Arab/West Asian	11 (2.7%)
Black	2 (0.5%)
Indigenous	8 (2.0%)
Latin American	8 (2.0%)
Mixed race	63 (15.4%)
Other	4 (1.0%)
HbA_1c at diagnosis, median [IQR]
%	11.7 [9.8, 13.7]
mmol/mol	104.4 [83.6, 126.2]
HbA_1c at study start (%), median [IQR]
%	7.8 [7.1, 8.6]
mmol/mol	61.7 [54.1, 70.5]
Total clinic visits, median [IQR]	5 [4,6]
Economic dependency quintile, n (%)
1 (least deprived)	152 (31.9%)
2	103 (21.6%)
3	96 (20.1%)
4	72 (15.1%)
5 (most deprived)	54 (11.3%)
Residential instability quintile, n (%)
1 (least deprived)	71 (14.9%)
2	129 (27.0%)
3	114 (23.9%)
4	96 (20.1%)
5 (most deprived)	67 (14.0%)
Situational vulnerability quintile, N (%)
1 (least deprived)	166 (34.8%)
2	116 (24.3%)
3	86 (18.0%)
4	78 (16.4%)
5 (most deprived)	31 (6.5%)

IQR, interquartile range; HbA_1c, hemoglobin A1c.

Across the three studied CIMD dimensions, there were no observed differences seen among the five levels of deprivation for HbA_1c and TIR (Fig. 1). Results were similar for HbA_1c and TIR after accounting for repeated measures and other covariates. For example, the mean difference in HbA_1c between the lowest (least deprived) and highest (most deprived) economic instability quintiles was 0.06 [95% CI = (−0.46, 0.58)]. After adjustment, the probability of pump use was similar across economic and situational dimensions, but those with the highest level of deprivation in the residential instability dimension had lower probability of pump use (e.g., −18.9%, 95% CI = −26.1% to −11.7% for quintile 5 vs. 1). See Table 2 for detailed adjusted estimates between quintiles. The most deprived quintiles of economic dependency and situational instability had the highest incidence rates of both DKA and severe hypoglycemia, whereas there was no clear pattern for residential instability (Table 3).

FIG. 1.

A1c, time in range, and use of pump by deprivation quintiles.

Table 2.

Relationship Between Deprivation Quintiles and A1c and Time in Range during the Study Period (June 10, 2020–December 7, 2022)

	Economic	Residential	Situational
HbA_1c (mmol/L), mean difference (95% CI)
Q1	Reference	Reference	Reference
Q2	0.17 (−0.25, 0.59)	0.34 (−0.15, 0.83)	0.22 (−0.17, 0.62)
Q3	−0.04 (−0.46, 0.39)	0.47 (−0.04, 0.97)	0.39 (−0.04, 0.81)
Q4	0.28 (−0.19, 0.75)	0.54 (0.01, 1.06)	0.23 (−0.22, 0.69)
Q5	0.06 (−0.46, 0.58)	0.34 (−0.22, 0.89)	0.81 (0.17, 1.44)
Time in range (%), mean difference (95% CI)
Q1	Reference	Reference	Reference
Q2	2.88 (−2.11, 7.88)	−4.16 (−9.90, 1.57)	2.22 (−2.63, 7.08)
Q3	0.01 (−5.24, 5.25)	−3.91 (−9.72, 1.89)	−1.45 (−6.57, 3.68)
Q4	2.56 (−3.09, 8.22)	0.06 (−6.42, 6.54)	0.65 (−4.86, 6.15)
Q5	−3.30 (−10.07, 3.46)	−2.76 (−9.65, 4.14)	−5.33 (−13.72, 3.06)
Pump use (%), mean difference (95% CI)
Q1	Reference	Reference	Reference
Q2	8.2 (2.8, 13.7)	−5.5 (−11.8, 0.9)	0.5 (−4.8, 5.7)
Q3	−7.2 (−12.5, −1.8)	−4.3 (−10.8, 2.2)	−0.6 (−6.3, 5.1)
Q4	−2.3 (−8.5, 3.9)	−8.0 (−14.9, −1.1)	−1.6 (−7.6, 4.3)
Q5	0.5 (−6.3, 7.3)	−18.9 (−26.1, −11.7)	−5.1 (−13.2, 3.0)

All data are presented as mean (for A1c and time in range) and risk (for pump use) differences and 95% confidence intervals adjusted for age at diagnosis, age at measurement, and sex.

95% CI, 95% confidence interval.

Table 3.

Counts and Incidence Rates of DKA and Hyperglycemia during the Study Period (June 10, 2020–December 7, 2022)

	Quintile 1	Quintile 2	Quintile 3	Quintile 4	Quintile 5
Recent DKA
Economic dependency	9/670 (1.3)	2/476 (0.4)	10/434 (2.3)	10/324 (3.1)	9/253 (3.6)
Residential instability	7/316 (2.2)	7/604 (1.2)	14/525 (2.7)	7/407 (1.7)	5/305 (1.6)
Situational vulnerability	14/761 (1.8)	3/503 (0.6)	7/408 (1.7)	10/344 (2.9)	6/141 (4.3)
Recent severe hypoglycemia
Economic dependency	9/670 (1.3)	21/476 (4.4)	12/434 (2.8)	13/324 (4.0)	16/253 (6.3)
Residential instability	8/316 (2.5)	20/604 (3.3)	22/525 (4.2)	11/407 (2.7)	10/305 (3.3)
Situational vulnerability	23/761 (3.0)	8/503 (1.6)	20/408 (4.9)	6/344 (1.7)	14/141 (9.9)

Denominators represent the number of visits.

The analysis population pre- and postuniversal coverage of CGM was similar across all demographics (Supplementary Table S3). Across each measure, those with higher levels of deprivation had lower rates of sensor use throughout the study period. For example, the risk difference between highest and lowest economic deprivation level was −16.6% (−23.3%, −9.8%) (Supplementary Table S4). After the introduction of universal CGM funding, there was an increase in CGM uptake among all levels of deprivation across all CIMD dimensions. The benefit was generally greatest among the most deprived groups (quintiles 4 and 5, P value for differential change between quintiles <0.001) (Fig. 2). The difference in sensor use from the most to least deprived situational group was −21.0% (−35.4%, −6.6%) at the time of universal coverage and shrank to −4.6% (−21.6%, 12.4%) after 12 months of coverage. A similar but more modest change was seen for the economic and residential components.

FIG. 2.

Estimated use of sensor pre- and post-Pharmacare (dotted line) coverage by deprivation quintiles. Solid lines represent fitted values from generalized additive model, including a segmentation at the time of Pharmacare coverage.

Discussion

Summary

In this Canadian cohort of 477 children and youth with T1D, we report an overall increase in CGM uptake across all levels of deprivation with the greatest increase observed in the most deprived groups 1 year following the implementation of a universal CGM funding program in the Canadian province of BC. Notably, differences in CGM use across indices of deprivation did not completely disappear 1 year after universal CGM coverage. Interestingly, for pump use where universal coverage has been in place for over a decade, we did not find significant differences across deprivation quintiles, except for the residential variability dimension where there was a difference only between the least and most deprived groups. We did not observe any differences across levels of deprivation in HbA_1c and TIR. Concerningly, the highest rates of DKA and episodes of severe hypoglycemia were in the most deprived groups for economic dependency and situational vulnerability, but overall rates were low, and results should be interpreted cautiously.

Interpretation and comparison with existing literature

In Australia, CGM uptake increased from 5% to 79% two years after initiation of a national public policy supporting universal coverage of CGM in people <21 years of age living with T1D.¹³ The 20% who did not access CGM via this national policy were older, had longer duration of diabetes, were more likely to be managed on basal/bolus regimens, and had a higher HbA_1c; however, the impact of SES disparities on nonuse was not described.¹³ Using the same national Australian registry, Lomax et al. looked at CGM use 5 years after universal CGM coverage was implemented in youth <18 years of age living with T1D and found that CGM use was similar across all socioeconomic groups except the most deprived quintile.²⁴ Data from the German Prospective Follow-up Registry showed that after implementation of public coverage of CGM in 2016, the equity gap in CGM use between the least and most deprived waned over time, and 3 years later (2019), the effect disappeared.²⁵ In our study, 1 year following universal CGM coverage, we continued to observe disparities in CGM use between the least and most deprived quintiles; however, more than a decade after a universal pump program was established in BC, a disparity in pump use persisted between the most and least deprived quintile groups only for the residential instability dimension. Future research will be important to establish whether the equity gap in CGM use in our region of Canada will disappear over time like in Germany or persist between the least and most deprived groups like in Australia.

Different from what has been reported in the literature, we found that HbA_1c did not differ across levels of social deprivation. In Australia’s national cohort of youth, HbA_1c was 0.51% higher in the most deprived compared with the least deprived groups.²⁴ In a Canadian population-based study from the province of Ontario, HbA_1c was highest in patients with the greatest degree of deprivation.⁸ Population studies from both England and Wales and New Zealand showed that children from minority ethnic groups and who experience high levels of deprivation had significantly higher HbA_1c compared with White children and the least deprived children, respectively.^26,27 In a transatlantic comparison between data from the U.S. T1D Exchange Registry and the German DPV Registry, HbA_1c was significantly higher in the most deprived compared with the least deprived groups; however, the magnitude of difference between deprivation groups was less pronounced in the German population.¹ It is possible that this dampened association in Germany compared with the United States may partly be due to Germany’s universal health care system.

Like in Germany, it may also be that Canada’s universal health care system dampens the impact of SES disparities on health outcomes (i.e., HbA_1c). In addition, some children and their families living with diabetes in the province of BC are followed by regional or community-based diabetes centers because the Diabetes Clinic at BCCH is in the southwest corner of the region making travel for some families lengthy and financially constraining. It is possible that in some of these centers, there is a disproportionate number of children with higher levels of deprivation. As the BC-PDR does not currently collect data on patients seen in regional or community-based clinics, our data may not sufficiently capture children from the most deprived quintiles.

One-year postuniversal CGM coverage, the overall rate of CGM use was still less than 50%. Therefore, describing CGM uptake over a longer time span than studied herein will determine whether the observed uptake further increases and is sustained over time and if the disparity gap in CGM uptake continues to diminish. Expanding our cohort to children accessing pediatric diabetes care in regional community-based clinics, work that is already underway, will improve the generalizability of our results.

Strengths and limitations

Study strengths include that our sample was reasonably large and our population was diverse particularly with respect to the economic dependency and residential instability quintiles. This study also has several limitations. First, the deprivation index quintiles are derived at the dissemination area level and, therefore, may not accurately represent all individuals living within the dissemination area. This is an inherit feature of the index, but creates the possibility of misclassification biases when applied at the individual level. Second, our dataset was derived exclusively from patients accessing care at the only tertiary care children’s hospital in BC. Our results on clinical outcomes stratified by deprivation index may not be generalizable to children accessing care in regional or community-based diabetes clinics in BC. Furthermore, children in the most deprived quintile may be more likely to reside in rural or remote areas that are served by community-based diabetes clinics, which may have affected our results by underestimating the number of children in the most deprived quintile. Third, at the time of implementation of universal CGM funding in BC, there was a required application process and variable waiting period of up to 2–3 months that may have impacted the degree of uptake seen in the first year. The process for applying for CGM coverage in BC has since improved via implementation of a streamlined online application process. In addition, during the study period, the clinical workflow likely had not fulsomely adopted processes where CGM was introduced to all patients, including those with a new diagnosis of T1D. Fourth, our study did not evaluate sustained use of CGM (i.e., for >1 year) by deprivation index, an important area for future research. Finally, our sample of patients experiencing DKA and hyperglycemia in the most deprived quintiles were low, limiting our ability to conduct adjusted analyses.

Conclusion

There is ample evidence showing that access to diabetes technology significantly improves health outcomes in all children and youth living with T1D. Yet, too many young people do not benefit because of widening inequities in care. Our study showed that the implementation of a universal CGM funding policy in the province of BC was effective in narrowing, but not eliminating, this gap. Our data are consistent with prior publications²⁸ that demonstrate that diabetes technology access is a first step, but does not negate structural drivers of inequity,⁹ which must be addressed to establish equity and justice. These data will inform future public policy decisions as new technology emerges, such as automated insulin delivery systems, which have no doubt been a boon in pediatric T1D management. The time is now to develop multipronged strategies so that the latest diabetes technologies are in the hands of all children and youth living with diabetes. Our study adds to the growing evidence that public coverage of diabetes technology is a key component of these strategies and is essential in moving the needle toward equity and justice in diabetes care.

Footnotes

Acknowledgments

Thank you to all of the patients who contributed their data to the BC Pediatric Diabetes Registry.

Authors’ Contributions

S.A.: Conceptualization, methodology, investigation, writing—original draft, visualization, supervision, and funding acquisition. J.B.: Methodology, investigation, writing—original draft, visualization, and formal analysis. C.L.: Methodology, visualization, and writing—review and editing. A.A.: Methodology, visualization, and writing—review and editing. S.A. 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 accuracy of the data analysis. The authors did not use artificial intelligence, machine learning, language models, or similar technology to create the content in this article or to assist with writing or editing the article.

Author Disclosure Statement

S.A. holds a BC Children’s Hospital Research Institute Research Salary award. S.A. has participated on advisory boards for Dexcom, Abbott, Novo Nordisk, Eli Lilly, Sanofi, and Insulet. No other potential conflicts of interest relevant to this article were reported.

Funding Information

This study and the BC Pediatric Diabetes Registry are funded through philanthropic funding from the BC Children’s Hospital Foundation. The funder had no role in the data collection and analysis, drafting of the article, or the decision to publish.

Supplementary Material

Supplementary Table S1

References

Addala

, Auzanneau

, Miller

, et al. A decade of disparities in diabetes technology use and HbA_1c in pediatric type 1 diabetes: A transatlantic comparison. Diabetes Care, 2021; 44(1):133–140; doi: 10.2337/dc20-0257

Lipman

, Hawkes

. Racial and socioeconomic disparities in pediatric type 1 diabetes: Time for a paradigm shift in approach. Diabetes Care, 2021; 44(1):14–16; doi: 10.2337/dci20-0048

Lipman

, Smith

, Patil

, et al. Racial disparities in treatment and outcomes of children with type 1 diabetes. Pediatr Diabetes, 2021; 22(2):241–248; doi: 10.1111/pedi.13139

Fegan‐Bohm

, Minard

, Anderson

, et al. Diabetes distress and HbA1c in racially/ethnically and socioeconomically diverse youth with type 1 diabetes. Pediatr Diabetes, 2020; 21(7):1362–1369; doi: 10.1111/pedi.13108

Addala

, Maahs

, Scheinker

, et al. Uninterrupted continuous glucose monitoring access is associated with a decrease in HbA_1c in youth with type 1 diabetes and public insurance. Pediatr Diabetes, 2020; 21(7):1301–1309; doi: 10.1111/pedi.13082

Senniappan

, Hine

, Tang

, et al. North West England Paediatric Diabetes Network. The effect of socioeconomic deprivation on efficacy of continuous subcutaneous insulin infusion: A retrospective paediatric case-controlled survey. Eur J Pediatr, 2012; 171(1):59–65; doi: 10.1007/s00431-011-1482-x

Agarwal

, Simmonds

, Myers

. The use of diabetes technology to address inequity in health outcomes: Limitations and opportunities. Curr Diab Rep, 2022; 22(7):275–281; doi: 10.1007/s11892-022-01470-3

Zuijdwijk

, Cuerden

, Mahmud

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

Hill-Briggs

, Adler

, Berkowitz

, et al. Social determinants of health and diabetes: A scientific review. Diabetes Care, 2020; 44(1):258–279; doi: 10.2337/dci20-0053

10.

Odugbesan

, Addala

, Nelson

, et al. Implicit racial–ethnic and insurance-mediated bias to recommending diabetes technology: Insights from T1D exchange multicenter pediatric and adult diabetes provider cohort. Diabetes Technol Ther, 2022; 24(9):619–627; doi: 10.1089/dia.2022.0042

11.

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

12.

Dicembrini

, Cosentino

, Monami

, et al. Effects of real-time continuous glucose monitoring in type 1 diabetes: A meta-analysis of randomized controlled trials. Acta Diabetol, 2021; 58(4):401–410; doi: 10.1007/s00592-020-01589-3

13.

Johnson

, Holmes-Walker

, Chee

, et al. ADDN Study Group. Universal subsidized continuous glucose monitoring funding for young people with type 1 diabetes: Uptake and outcomes over 2 years, a Population-Based Study. Diabetes Care, 2022; 45(2):391–397; doi: 10.2337/dc21-1666

14.

DeSalvo

, Lanzinger

, Noor

, et al. Transatlantic comparison of pediatric continuous glucose monitoring use in the Diabetes-Patienten-Verlaufsdokumentation initiative and type 1 diabetes exchange quality improvement collaborative. Diabetes Technol Ther, 2022; 24(12):920–924; doi: 10.1089/dia.2022.0248

15.

Lanzinger

, Best

, Bergmann

, et al. Dynamics of hemoglobin A1c, body mass index, and rates of severe hypoglycemia in 4434 Adults with type 1 or type 2 diabetes after initiation of continuous glucose monitoring. Diabetes Technol Ther, 2022; 24(10):763–769; doi: 10.1089/dia.2022.0063

16.

Franceschi

, Pertile

, Marigliano

, et al. Satisfaction with continuous glucose monitoring is associated with quality of life in young people with type 1 diabetes regardless of metabolic control and treatment type. Diabet Med, 2024; 41(6):e15307; doi: 10.1111/dme.15307

17.

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

18.

Ayub

, Ng

, Portales-Casamar

, et al. Toward building a provincial diabetes registry of children and youth living with diabetes in British Columbia, Canada. Can J Diabetes, 2022; 46(4):346–352.e1; doi: 10.1016/j.jcjd.2021.11.006

19.

Matheson

, Dunn

, Smith

KLW

, et al. Development of the Canadian Marginalization Index: A new tool for the study of inequality. Can J Public Heal Rev Can Sante Publique, 2011; 103(8 Suppl 2):S12–S6.

20.

Bernal

, Cummins

, Gasparrini

. Interrupted time series regression for the evaluation of public health interventions: A tutorial. Int J Epidemiol, 2017; 46(1):348–355; doi: 10.1093/ije/dyw098

21.

Naimi

, Whitcomb

. Estimating risk ratios and risk differences using regression. Am J Epidemiol, 2020; 189(6):508–510; doi: 10.1093/aje/kwaa044

22.

Wood

. Generalized Additive Models, 2nd ed.. Chapnman and Hall: New York; 2017; doi: 10.1201/9781315370279

23.

The R Project for Statistical Computing. Available from: https://www.R-project.org/ [Last accessed: July 30, 2024 ].

24.

Lomax

, Taplin

, Abraham

, et al. Socioeconomic status and diabetes technology use in youth with type 1 diabetes: A comparison of two funding models. Front Endocrinol (Lausanne), 2023; 14:1178958; doi: 10.3389/fendo.2023.1178958

25.

Auzanneau

, Rosenbauer

, Maier

, et al. Heterogeneity of access to diabetes technology depending on area deprivation and demographics between 2016 and 2019 in Germany. J Diabetes Sci Technol, 2021; 15(5):1059–1068; doi: 10.1177/19322968211028608

26.

Khanolkar

, Amin

, Taylor‐Robinson

, et al. Young people with Type 1 diabetes of non‐white ethnicity and lower socio‐economic status have poorer glycaemic control in England and Wales. Diabet Med, 2016; 33(11):1508–1515; doi: 10.1111/dme.13079

27.

Carter

, Cutfield

, Hofman

, et al. Ethnicity and social deprivation independently influence metabolic control in children with type 1 diabetes. Diabetologia, 2008; 51(10):1835–1842; doi: 10.1007/s00125-008-1106-9

28.

Addala

, Ding

, Zaharieva

, et al. Teamwork, Targets, Technology, and Tight Control (4T) Study Group. Disparities in hemoglobin A1c levels in the first year after diagnosis among youths with type 1 diabetes offered continuous glucose monitoring. JAMA Netw Open, 2023; 6(4):e238881; doi: 10.1001/jamanetworkopen.2023.8881

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

0.08 MB