Demographic Variables Associated With Diabetes Technology Awareness or Use in Adults With Type 2 Diabetes

Eligibility criteria included age ≥40 years, diagnosis of type 2 diabetes (as identified by International Classification of Diseases, 10th revision, codes beginning with E11), treatment with insulin, English as a primary language, and documentation of an e-mail address in the electronic health record (EHR). Eligible patients were recruited from four primary care practices in New York’s Queens and Nassau counties; two were attending-only practices, one was an attending-resident hybrid practice, and one was a resident-only practice. Three of the four practices have a certified diabetes care and education specialist. To be included in this study, patients had to have had at least one visit to one of the four practices between 1 July 2018 and 30 June 2021.

This study was approved by Northwell’s institutional review board (IRB #21-0632). Patients identified in the EHR were emailed to participate using a de-identified link to a questionnaire in REDCap (Research Electronic Data Capture) (10,11). After clicking the link, patients were first presented with a written informed consent form and, if they agreed, they would proceed to the questionnaire. Seven hundred and thirty patients met the eligibility criteria and were emailed invitations to participate between 7 August and 5 October 2022. Each patient was emailed a total of six times and reminded not to answer the questionnaire more than once. Fifty-nine e-mail addresses were found to be invalid. One hundred and thirty-five respondents were collected, of which 17 did not consent to participate and were excluded, along with two patients with type 1 diabetes. The final analysis included 116 participants.

The survey solicited demographic variables, including age, sex, race, ethnicity (Hispanic vs. non-Hispanic), medical insurance type, highest level of education, primary language, and annual household income. The survey also collected information regarding diabetes history, including type and duration of diabetes, diabetes provider type, and most recent A1C. Participants were then asked about their awareness and use of insulin pumps and CGM.

The sample size for this study was based on the availability of subjects (i.e., a convenience sample) as well as feasibility of recruitment and was exploratory in nature. Participant awareness and use of CGM and insulin pumps were summarized descriptively using frequencies and percentages and were compared by racial/ethnic group, education level, and annual household income using Pearson χ² or Fisher exact tests. Racial/ethnic minority group and annual household income were collapsed to facilitate analysis (Hispanic or non-White vs. non-Hispanic and White, and <$25,000 vs. ≥$25,000 per year, respectively). Education level was first collapsed into high school or less versus more than high school and second by high school or less versus some college versus graduate school. All results were considered statistically significant at the P <0.05 level of significance. Analyses were performed using R, v. 4.1.2, statistical software.

Participant characteristics are summarized in Table 1. Most participants were ≥65 years of age (58%), White (62%), female (52%), and Medicare recipients (56%). Most participants had diabetes for >15 years (61%), had a most recent A1C of 7–9% (62%), and received diabetes care from an endocrinologist (69%).

Participants were divided into two racial/ethnic categories: non-Hispanic White (n = 69) and racially/ethnically minoritized (n = 46) (Table 2). Awareness of CGM, insulin pumps, and either CGM or insulin pumps was found to be significantly lower in the racially/ethnically minoritized group compared with the non-Hispanic White group (72 vs. 90%, P = 0.013; 70 vs. 92%, P = 0.001; and 85 vs. 99%, P = 0.007, respectively) (Table 2). However, among those with awareness of insulin pump therapy, current insulin pump use was higher in the racially/ethnically minoritized group compared with the non-Hispanic White group (25 vs. 5%, P = 0.007) (Table 2).

Participants were first divided into two education categories: high school or less (n = 13) and more than high school (n = 103) (Table 3). Awareness of insulin pump therapy was found to be significantly lower in participants with less education (61 vs. 86%, P = 0.041) (Table 3). However, among those with awareness of insulin pump therapy, current insulin pump use was higher in participants with high school or less education (38 vs. 9%, P = 0.048) (Table 3).

When patients were divided into three education categories: high school or less versus some college versus graduate school, there was no difference in awareness of CGM or insulin pumps across different education levels. Among patients with awareness of insulin pumps, the rate of current insulin pump use was found to be highest in those with a high school education or less, intermediate in those with some college education, and lowest in those with graduate school education (38, 15, and 0%, respectively; P = 0.003).

Among the 95 participants (82%) who chose to report their income, neither awareness nor use of insulin pumps and/or CGM was found to be associated with income (Table 4).

This study found that awareness of both CGM and insulin pumps was lower in the racially/ethnically minoritized group, and awareness of insulin pump therapy was lower in the less educated group. On the other hand, among those with awareness of insulin pump therapy, current insulin pump use was higher in the racially/ethnically minorized group and the less educated group. Notably, neither awareness nor use of insulin pumps and/or CGM was found to be associated with income. These findings differ from studies of disparities in type 1 diabetes, which consistently show lower use of insulin pumps and CGM in patients of racially/ethnically minoritized groups and lower socioeconomic status (1–3).

Health care providers’ implicit biases and other preconceptions may play an important role in the disparity in awareness of diabetes technology. Numerous studies point out that health care providers have implicit biases toward racially/ethnically minoritized groups, and these biases negatively affect patient-provider interactions and treatment outcomes (12,13). This study was limited by its ability to assess this factor, and measuring provider bias should be considered for future studies.

Several studies have shown that providers only discuss diabetes technology with patients they perceive will use it. In a study of providers for children and teens with type 1 diabetes, family support and high levels of education were deemed the positive characteristics meriting referral of patients for a closed-loop system (14). Over time, that view changed as providers realized that some highly educated patients would manipulate their pumps, whereas those with lower education levels were more willing to follow the instructions (14). These biases can be overcome by engaging providers in insulin pump studies that randomize all comers to using a pump or a multiple daily injection insulin regimen. Providers can overcome biases against older or unmotivated patients by seeing them improve with insulin pump use (15).

This study has several limitations. First, this is a cross-sectional study and thus it cannot be used to test causal relationships among variables. Second, a series of logistic regression analyses looking at the relationship between the main exposures (race/ethnicity, education, and annual income) and outcomes (CGM and insulin pump awareness and use) adjusted for any potential confounders of interest (e.g., age, sex, and type of diabetes care provider) would strengthen the study. Unfortunately, multivariable logistic regression requires a minimum of 10 observations per outcome group for each covariate included in the model, which this study lacked. Third, the four different practices in this study all belong to one health system in the New York metropolitan area. The majority of participants (99%) had health insurance, and 69% had access to an endocrinologist; thus, the results may not be generalizable. Fourth, this study distributed questionnaires via e-mail, and thus participants were more likely to be technologically savvy, which may correlate with their awareness and use of diabetes technology, leading to a selection bias. Fifth, study participants were mostly non-Hispanic White and Black individuals, with only 3% of Hispanic ethnicity and <11% identifying as Asian or other races. The requirement of English as a primary language could have led to the exclusion of Hispanic or Asian patients who speak English as a second language.

A major strength of this study is that it was performed in the primary care setting, where most people with type 2 diabetes are treated (16). Given the increased availability of diabetes technology, primary care physicians play an important role in introducing and managing these devices (17,18). Type 2 diabetes disproportionately affects people in racially/ethnically minoritized groups and those with lower education levels and lower socioeconomic status (16), but there is a paucity of research investigating disparities in diabetes technology in type 2 diabetes in primary care settings (19). Although the majority of participants in this study see an endocrinologist, this finding is not generalizable to the general population. Healy et al. (20) noted that, in 2015, there were 15.5 adult endocrinologists per 1 million people (20). Considering that there are >30 million people with diabetes, many will have to rely on their primary care physicians for care.

Future studies need to control for more socioeconomic variables and have a more racially/ethnically and socioeconomically diverse patient base.

Among a sample of outpatients with type 2 diabetes, those who identified as racial/ethnic minorities and those with less than a high school education were less likely to have awareness of insulin pumps and CGM. In addition, providers need to be aware of their own implicit biases when prescribing or discussing diabetes technology with their patients. More investigations are needed to evaluate the causes of these disparities and improve access to diabetes technology for racially/ethnically minoritized and socioeconomically disadvantaged groups.

No potential conflicts of interest relevant to this article were reported.

Y.Y. and B.A.A.M. collected data and wrote the manuscript. S.I. analyzed the data. S.I. and A.K.M. contributed to the discussion and reviewed/edited the manuscript. Y.Y. 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.