Recent advances in diabetes technology have transformed the management of Type 1 diabetes (T1D) leading to improvements in glycemic control and quality of life. Randomized controlled trials have demonstrated that automated insulin delivery (AID) systems – consisting of a continuous glucose monitor (CGM), insulin pump and control algorithm that adjusts insulin delivery based on glucose levels – lead to improvement in A1c and reduction in hypoglycemia in adults and children with T1D (1, 2, 3, 4, 5). The American Diabetes Association Professional Practice Committee recommends that AID systems be offered for diabetes management for youth and adults with T1D who can use the devices safely (6). To date, most of the randomized controlled trials evaluating these technologies in T1D have been in more socially advantaged and predominantly non-Hispanic White patients followed at specialized diabetes care centers (1, 2, 3, 4). Furthermore, these efficacy studies have largely excluded patients with T1D who have markedly elevated A1c levels.

Although individuals with T1D from minoritized communities have poorer health outcomes, use of diabetes technologies in these populations remains low (7, 8, 9, 10). Some of the more common reasons for these inequities include well-recognized social determinants of health, including barriers to access to care, health insurance coverage limitations, limited health literacy, and that racial and ethnic minority populations have largely been excluded from many of the randomized controlled trials evaluating diabetes technologies. Recently, clinician implicit bias has been identified as an important contributor to poor device uptake. Providers who prescribe diabetes technologies are highly influential and act as “gatekeepers” for their patients, making specific device recommendations, supporting device initiation, and providing longitudinal support (11). However, many adult endocrinology trainees feel under-prepared in critical aspects of technology use, creating an additional potential barrier to device uptake (12). Moreover, even in healthcare systems with universal insurance coverage for insulin pumps and CGM, this has not necessarily resolved these disparities (13).

In addition to these nationwide issues previously mentioned, Boston Medical Center (BMC), the largest safety-net hospital in New England, faces additional local challenges. Our specialty diabetes clinic is an adult referral clinic located within the Department of Medicine and Section of Endocrinology, Diabetes and Nutrition. Patients from minoritized groups, who are public insurance beneficiaries and those from racial and ethnic minority groups, constitute over 60% of our population and one third do not speak English as their primary language. Our diabetes program serves a majority type 2 diabetes population, the vast majority of which are not using AID systems, with a much smaller population of adults with T1D (approximately 500 patients, with 70-90 patient visits per month). This has resulted in limited advanced diabetes technology expertise amongst our clinicians who are not working in a high-volume practice with many patients using AID systems. Furthermore, we have also reported disparities in AID use within our own clinic population, in which Black and Hispanic patients, despite similar insurance coverage, diabetes duration, educational level, and prior pump use, had worse clinical outcomes (14). Technology also advances quickly, which may result in clinician discomfort with recommending newer devices. Finally, many of the patients entering our diabetes program have limited diabetes self-management skills and nutritional literacy.

Our adult diabetes clinic joined the Type 1 Diabetes Quality Improvement Collaborative (T1DX-QI) in early 2021. This nationwide learning collaborative comprises over 50 US-based academic diabetes centers who share, and benchmark, clinical data derived from over 80,000 people living with T1D. Most of these centers serve predominantly commercially-insured, white populations and pediatric diabetes centers comprise the bulk of the membership. Of the adult centers, less than 10 are regarded as safety-net institutions. The T1DX-QI facilitates clinical collaboration and collective learning and it provides funding support for QI initiatives to improve outcomes. After joining the T1DX-QI and benchmarking our data with US based peers, we became aware that use of CGM in conjunction with multiple daily injections in our adult T1D population, exceeded national benchmarks for CGM use (80% in our population). However, despite the absence of insurance barriers to diabetes technology use by Medicaid insured patients in Massachusetts, we recognized in the summer of 2021 that <15% of our T1D population were using insulin pumps/AID systems. This contrasted with the 60-66% insulin pump/AID system use reported by the adult population in the T1D Exchange Clinic Registry in 2016-2018 (15).

In the fall of 2021, our diabetes clinic initiated a transformative program aimed at increasing uptake and reducing inequities in use of AID systems in our T1D population. We established a multi-disciplinary clinical team and applied quality improvement principles, including multiple iterative Plan-Do-Study-Act (PDSA) cycles, to identify barriers and develop solutions. We placed a particular focus on de-emphasizing “ideal” insulin pump candidacy and overcoming clinical inertia to referral of patients with T1D to a revamped insulin pump/AID educational program. We also placed emphasis on engaging and enrolling patients with A1c levels of >8.5% in pump education.

Our primary aim was to expand the use of AID systems by patients with T1D followed in our diabetes clinic. Central to our goal of broadening access, we loosened the criteria for AID eligibility to include patients with markedly elevated A1c levels, a patient group that has been excluded from the AID efficacy trials. We set an initial 1-year goal to increase AID system use rates by 30% over baseline.