Strategies to Increase the Use of Connected Insulin Pens Through Shared Decision-Making: A Quality Improvement Success Story

The Northwestern Academic Endocrinology practice, located in an urban setting in downtown Chicago, IL, is made up of 11 attending physicians, four clinical fellows, one physician assistant (PA), two nurse practitioners, and two certified diabetes care and education specialists (CDCESs). The practice serves a racially and ethnically diverse adult population with various levels of socioeconomic status and insurance coverage. We are affiliated with McGaw Medical Center and Northwestern University Feinberg School of Medicine and thus serve as a teaching practice for residents, medical students, PA students, and other learners, in addition to our fellowship training. We are part of the T1D Exchange Quality Improvement Collaborative (T1DX-QI), a network of pediatric and adult clinical teams serving patients with diabetes and sharing best practices for optimal and equitable care delivery through the use of quality improvement (QI) methodology and techniques (1).

Although our practice treats and manages a wide range of endocrine-related conditions, a majority of care is focused on the management of type 1 diabetes, type 2 diabetes, and other forms of diabetes. The practice is the home of a population health–focused “Diabetes Tune-Up” pathway, through which primary care providers within the Northwestern Medicine health care system can work together with specialized diabetes care providers to co-manage and promote the self-management skills in individuals with poorly controlled, high-risk diabetes.

Connected insulin pens (CIPs) are reusable, injectable insulin delivery devices that are connected via Bluetooth to a smartphone app and specifically designed to assist with insulin dose calculation, dosing recommendations, and tracking of dose delivery data (2). These pens, of which there are currently several types on the market made by various manufacturers, can provide a way for individuals with diabetes to better engage with their diabetes self-management, with easier tracking of doses and therapy adherence, improved data access, and recording of data for later review. Additionally, CIPs can assist in minimizing the risk of hypoglycemia from overcorrection of hyperglycemia because they can track users’ insulin on board and make specific correction dose recommendations taking this information into account.

While functionally similar to traditional insulin pens, CIPs offer the benefits of connectivity, providing data to both people with diabetes and their providers to enhance diabetes care and work toward improved outcomes. A study of individuals with type 1 diabetes who started using a CIP found improvements in continuous glucose monitoring (CGM) metrics such as time in range and time below range (<54 mg/dL) and fewer missed bolus doses (3).

Despite the benefits of CIPs, provider surveys have demonstrated that barriers to their use persist, limiting successful CIP prescribing. These barriers include lack of provider awareness and training, cost, and variable insurance coverage (4). Additionally, and of concern to our team with regard to diabetes prescribing practices, there are major disparities in prescribing and use of diabetes technology among marginalized populations, and these disparities result in widening inequities in diabetes-related complications and outcomes (5).

Our initiative aimed to address inequities in diabetes technology access and use in our clinic population by targeting all individuals with diabetes who were on multiple daily injection (MDI) insulin therapy and discussing CIP technology with them through a shared decision-making approach. This project was completed alongside other T1DX-QI sites as a means of sharing best practices to improve care for all people with diabetes. The T1DX-QI has a proven track record of improving care outcomes (6), and our hope was that this project would meaningfully affect and improve the lives of patients at all participating sites.

We identified this quality gap initially through anecdotal observations and discussions with providers at our site, as well as with other members of the T1DX-QI. There appeared to be a baseline lack of awareness of CIPs among both providers and people with diabetes that limited the ability to successfully prescribe and use this technology.

A total of 561 people with type 1 diabetes and 1,215 people with type 2 diabetes on MDI insulin therapy had a planned clinic visit during the study window and could have been eligible for a CIP.

This QI project started in June 2022 initially for patients with type 1 diabetes and expanded to include those with type 2 diabetes in November 2022. The project was completed in March 2023.

The project leader was a senior faculty physician with specific clinical and research interest in diabetes technology. She was selected for this role because of her wealth of knowledge and experience; she played a crucial role in helping to provide education to providers, trainees, and people with diabetes on the many benefits of diabetes technology in improving clinical outcomes and quality of life. A clinical fellow also played a large role in the project, with duties that included screening patients for eligibility, conducting provider outreach, tracking data, and exploring avenues for further improvement with a focus on individuals with type 1 diabetes. The team additionally included a research project coordinator, who performed similar tasks for individuals with type 2 diabetes. The other clinicians in the practice played a role in educating people with diabetes and prescribing CIPs, and the medical assistant team helped to administer and collect Diabetes Technology Attitudes (DTA) questionnaires (Supplementary Figure S1). These surveys served as a way to encourage discussion about diabetes technology between people with diabetes and providers and to promote shared decision-making. Additionally, the CDCESs provided hands-on learning and education for people with diabetes who were prescribed a CIP, with ongoing check-in to ensure that they were using their new device effectively.

The project was supported by the T1DX-QI, and thus our team extends beyond our local site to other member sites of the T1DX-QI, as well as its central support staff. We collectively held monthly meetings as a means of updating everyone on process maps (Supplementary Figure S2), project progress, and Plan, Do, Study, Act (PDSA) cycles. The relationship with the T1DX-QI has allowed our site to participate in uniform data collection and analysis in a way that informs the implementation of real-time change to clinical practice with a goal of advancing diabetes care (7).

We instituted monthly QI check-in meetings at our site and provided a mid-project online survey for clinic providers and staff to provide feedback. We also enhanced our internal data tracking for diabetes technology prescribing. Overall, with our experience participating in this project, we feel more prepared and comfortable with the idea of using QI methods as a means of bettering our practice.

We found that standardizing a process of alerting providers of potential candidates for a CIP (i.e., people with diabetes who were already on MDI insulin therapy) through a weekly e-mail message was a way to prime the providers for a shared decision-making discussion on CIPs with the individuals with diabetes who they would be seeing that week. Carrying out this process in a way that was inclusive of all patients who met these criteria was a way to minimize any biases in prescribing practices for diabetes technology. Given the existing disparities in diabetes technology use, we felt that standardization of this pre-screening process was extremely important because provider biases can be a barrier to recommending diabetes technology (8).

We also provided written, easily accessible instructions for prescribing CIPs to all clinic providers, as well as dot phrases (pre-formed blocks of text for clinical documentation entered using a keyboard shortcut) for the electronic health record (EHR) system to assist with documentation of shared decision-making about CIPs. Additionally, we included the DTA questionnaire as a part of our standard workflow for encouraging thoughtful discussion on diabetes technology, with important inclusivity of medical assistants in the administration and encouragement of questionnaire completion.

A major limitation in our efforts was the time needed to have thorough conversations about CIPs and diabetes technology discussions within the constraints of short clinic visits. There was also no formal assessment to ensure that individuals with diabetes comprehended the discussed potential benefits of CIPs before deciding whether they were interested in trying one. Other limitations were variability in insurance coverage for CIPs and a limited volume of insulin units that can be delivered at once by CIP devices. The limited possible dose volume was sometimes insufficient for bolus insulin doses for people with high insulin requirements. Although providers were encouraged to document the reasons why individuals with diabetes chose to decline a CIP, these data were not formally collected and analyzed for this project.

• Plan. All providers seeing people with type 1 diabetes were emailed a how-to guide for prescribing CIPs, with an overview of the components that need to be prescribed (e.g., CIP and insulin cartridges), step-by-step guidance for writing an electronic prescription, and troubleshooting tips.

• Do. We did this on a weekly basis for 4 weeks.

• Study. The number of prescribed CIPs for people with type 1 diabetes using MDI insulin therapy increased from one to three per week over the course of 4 weeks.

• Act. We continued to include the how-to guide in weekly emails to clinic providers, expanded to also send it to providers seeing people with type 2 diabetes on MDI insulin therapy, and refined the content of the guide based on provider feedback.

During the study period, there were 561 eligible people with type 1 diabetes who were on MDI insulin therapy and had a visit during the study period, and, of these, 79 (14.0%) were prescribed a CIP. There were also 1,215 eligible people with type 2 diabetes, and, of these, 19 (1.6%) were prescribed a CIP.

A total of 109 DTA questionnaires were completed and collected (52 from people with type 1 diabetes and 57 from those with type 2 diabetes). Of the 52 people with type 1 diabetes who completed a questionnaire, 11 (21.1%) were prescribed a CIP, and 41 were not. Of the 41 who were not prescribed a CIP, 6 expressed interest in an insulin pump and were started on our pump pathway. Of the 57 people with type 2 diabetes who completed a questionnaire, 4 (7%) were prescribed a CIP, and 53 were not.

The racial/ethnic demographics of all 79 people with type 1 diabetes who were prescribed a CIP during the study period were 5 (6.3%) Asian, 8 (10.1%) Hispanic, 3 (3.8%) non-Hispanic Black, 49 (62.0%) non-Hispanic White, 10 (12.6%) not reported, and 4 (5.1%) other. This compares to the total clinic population of people with type 1 diabetes in MDI insulin therapy, which is 3.6% Asian, 7.8% Hispanic, 13.5% non-Hispanic Black, 60.9% non-Hispanic White, 8.7% not reported, and 5.3% other.

The racial/ethnic demographics of all 19 people with type 2 diabetes who were prescribed a CIP during the study period were 3 (15.7%) Asian, 1 (5.3%) Hispanic, 8 (42.1%) non-Hispanic Black, 4 (21.1%) non-Hispanic White, 1 (5.3%) not reported, and 2 (10.5%) other. This compares to the total clinic population of people with type 2 diabetes on MDI insulin therapy, which is 7.1% Asian, 14.5% Hispanic, 22.5% non-Hispanic Black, 49.0% non-Hispanic White, 3.9% not reported, and 3.1% other.

As a site that is relatively new to QI, we had a very positive experience and plan to continue to use QI methodology in the future and to take the knowledge gained from this project (through both the ideas that worked well and those that required reevaluation) to further expand diabetes technology prescribing practices beyond just CIPs. We hope to continue to work toward a process in which a standardized, equitable workflow for shared decision-making on diabetes technology is incorporated for any clinic visits with people with diabetes who are not presently using diabetes technology. We plan to further explore ways to optimally use the technology questionnaires as a targeted tool for improving diabetes technology prescribing practices and utilization and explore more strategies to improve equity in diabetes technology use.

Specific to CIP use, we would also like to engage with our providers to assess their satisfaction with using CIPs in their practice and how the technology could be used to optimize their workflow. Additionally, although we did not collect data on A1C or CGM metrics before and after CIP initiation in this project, we hope to incorporate these factors into future work to assess clinical outcomes alongside data on diabetes technology uptake and utilization.

We found that we were able to increase CIP prescribing at our practice by using QI methodology and sharing best practices with other T1DX-QI sites. Through a multitude of changes, including using diabetes technology questionnaires, equitable screening practices, and sharing EHR dot phrases, we quickly implemented wider use of CIPs in our practice pattern. Of note, it did appear that CIP prescribing was higher for people with type 1 diabetes than for those with type 2 diabetes. We hope to further explore reasons for this finding and address barriers to CIP use in people with type 2 diabetes. Most of the interventions used in this project, including using the simple paper questionnaires and standardizing a diabetes technology conversation with all individuals with diabetes on MDI insulin therapy, could be adapted readily by other clinical sites regardless of practice size or resources.

When successful, QI work can be fantastically rewarding. However, to get to achieve meaningful change, it is crucial to incorporate the opinions and feedback of all members of a clinic environment. The diversity in viewpoints and experiences of team members is essential to a successful process.

The authors thank the Leona M. and Harry B. Helmsley Charitable Trust for funding the T1DX-QI. The authors also acknowledge the contributions of people with diabetes, families, diabetes care teams, and collaborators within the T1DX-QI who continually seek to improve care and outcomes for people living with diabetes.

This study was funded by an education grant from the Eli Lilly & Company Global Medical Affairs Division, Indianapolis, IN. The T1DX-QI is funded by the Leona M. and Harry B. Helmsley Charitable Trust.

G.A. has received research support to Northwestern University from Fractyl Health, Insulet, MannKind, Tandem Diabetes, and Welldoc and has served as a consultant for Dexcom, Insulet, and Medscape. No other potential conflicts of interest relevant to this article were reported.

J.G.F. assisted with process mapping, PDSA cycle idea generation and execution, subject screening, provider and staff outreach, data collection and tracking, and manuscript writing and revision. P.S. assisted with data tracking and technical/statistical support. S.H. assisted with subject screening, provider outreaching, and data tracking. G.A. assisted with process mapping, generated and executed PDSA cycles, assisted with data tracking, encouraged provider participation, facilitated shared decision-making, and reviewed and revised the manuscript. G.A. is the guarantor of this work and, as such, had full access to the data reported and takes responsibility for the integrity of the data and the accuracy of the data analysis.