Abstract #1 Interventions to improve technology equity in young adults (YA) with type 1 diabetes (TID)
Priyanka Mathias1 and Shivani Agarwal1,2
1Fleischer Institute for Diabetes and Metabolism, Albert Einstein College of Medicine, Montefiore Medical Center, Bronx, New York, USA; 2NY‐Regional Center for Diabetes Translational Research, Albert Einstein College of Medicine, Montefiore Medical Center, Bronx, New York, USA
pmathias@montefiore.org
Background: Diabetes technology has been shown to improve glycemic control in people with type 1 diabetes (T1D). However, young adults (YA) underutilize both continuous glucose monitors (CGM) and insulin pumps (IP) compared with other age groups. We examined whether a specialty YA T1D clinical program could improve rates of technology use over time.
Methods: The Supporting Emerging Adults with Diabetes program was developed in 2019 in an underserved medical center in New York to deliver multidisciplinary technology forward care to YA with T1D. The program care model includes care coordination, orientation to adult healthcare, continuing education, behavioral support, social needs management, and pediatric partnership. As a participating site of the T1D Exchange Quality Improvement Collaborative, we collected monthly aggregate data on CGM and IP prescription rates, with the numerator representing the number of people prescribed CGM/IP in the reporting month and the denominator representing the number of patients seen by endocrinology that month.
Results: In total, 253 YA were included (mean age: 23 years, 53% F, 55% Hispanic, 22% Non‐Hispanic Black [NHB], 10% White). Overall, CGM prescription rates increased from 31% to 69% from January 2019 to December 2021 (p < 0.001). Equal improvement was seen in Hispanic (12.5% to 71%) and NHB (33% to 55%), compared with White (33% to 75%). IP prescriptions improved from 6.3% to 31% overall [Hispanic (0% to 39%), NHB (0% to 37%), White (0% to 17%)].
Conclusions: A tailored program for YA with T1D can improve equity in technology uptake. Future studies are needed to examine the long‐term impact on care engagement measures and patient‐reported outcomes.
Keywords: continuous glucose monitor, diabetes technology, insulin pump, equity, young adults
Abstract #2 Improving back‐up planning in the event of pump failure
Victoria Elliott; Meghan Pauley; Erin Finn; Anna Valentine; Lauren Waterman; Rachel Sewell; Olivia Docter; Jennifer Barker; G. Todd Alonso
University of Colorado, Barbara Davis Center for Diabetes and the Children's Hospital of Colorado Aurora, Colorado, USA
victoria.elliott@childrenscolorado.org
Background/Objective: We identified that many pump users did not have active long‐acting insulin prescriptions, which complicates care in the event of pump failure. The aim of our quality improvement (QI) initiative was to improve planning for pump failure by increasing the percentage of insulin pump users who have active long‐acting insulin prescriptions.
Methods: Between August 2021 and March 2022, a multidisciplinary team of pediatric endocrinology fellows, faculty, nurses, and a medical record analyst performed seven Plan‐Do‐Study‐Act cycles. These cycles defined the problem, agreed on metrics and goals, solicited ideas from clinic staff and patients, and implemented a series of process changes. The most important intervention was adding standard language in the visit summary, prompting providers to provide long‐acting insulin dose instructions. We extracted data from the medical record to determine baseline and monthly prescription rates.
Results: At baseline, 54% of pump users seen between February 20, 2021 and August 20, 2021 had active long‐acting insulin prescriptions. After the intervention, the percentage of pump users with long‐acting insulin prescriptions was 70% in May 2022. There was a significant increase in active prescriptions when comparing baseline, planning, and post‐intervention time periods (p = <0.01, Figure 1).
Conclusions: Our clinic increased the rate of active long‐acting insulin prescriptions for pump users to improve planning for pump failure. We also provided written long‐acting insulin dose recommendations, which many families were unable to determine themselves. Future initiatives include improving patient knowledge of transitioning to injections in the event of pump failure.
Keywords: insulin pump, pediatric, type 1 diabetes
Figure 1
Abstract #3 Increasing accessibility to CGM in an equitable fashion
Blake Adams; Jayme Wasson; Kathryn Sumpter; Grace Nelson
University of Tennessee Health Science Center, Le Bonheur Children's Hospital Memphis, Tennessee, USA
blake.adams@lebonheur.org
Background/Objective: Le Bonheur Diabetes Clinic partners with families and aids them in finding the best ways to manage their diabetes. Continuous glucose monitoring (CGM) has improved safety outcomes, allowing users to respond to blood glucose trends in real‐time. Non‐Hispanic Black patients had lower usage compared to Non‐Hispanic White patients. We aim to reduce inequities in CGM use between Non‐Hispanic (NH) White and Non‐Hispanic Black type 1 diabetes (T1D) patients by increasing CGM use in NH‐Black patients from 20% to at least 30% by June 2022.
Methods: Our initial barrier was CGM insurance approvals. In response, we trained a CGM champion to streamline processes. She built relationships with DME companies and communicated with insurance companies. We created patient education folders for families and worked with staff to increase their knowledge of CGM. We increased EMR functionality, made orders trackable, and created a smart text to standardize documentation.
Results: Baseline data showed that 22% of NH‐Black patients used CGM regularly. As of spring 2022, we now have 28% of NH‐Black patients using CGM. Our overall clinic CGM use has increased from 52% to 62% in this same timeframe.
Conclusions: We have seen improvement in our percentage of Non‐Hispanic Black patients and overall patient populations that are utilizing the CGM. We continue to improve our processes and are increasing CGM usage.
Keywords: CGM, diabetes technology, health equity
Abstract #4 Towards risk‐based management of type 1 diabetes (T1D): Developing a population health dashboard based on performing diabetes self‐management habits
Craig Vandervelden; Brent Lockee; Susana R. Patton; Ryan McDonough; Joyce M Lee; Mark Clements
Children's Mercy Hospitals Kansas City, Missouri, USA
cavandervedlen@cmh.edu
Background/Objectives: Recent research on six pediatric type 1 diabetes (T1D) self‐management habits [J. M. Lee et al., JAMA Netw Open. 2021;4(10)], aka “the 6 habits”, which are commonly documented in the electronic health record, offers a new evidence‐based measure of engagement with self‐management. We created an interactive population health management tool that displays data related to “the 6 habits” for providers and identified a new habit that associates with glycemic control.
Methods: Diabetes device data and electronic medical records for 4293 youth ages 1–27 with T1D from a pediatric diabetes care network in the Midwest USA were imported into “D‐Data Dock,” a novel Microsoft Azure‐based data lakehouse developed by our team. We designed population health dashboards to display performance of “the 6 habits” plus a 7th habit (“consumes a healthy diet”) by youth, along with their relationship to HbA1c% and time in range (TIR). Providers can explore habit performance in the population by various demographics and drill down to the individual patient level.
Results: Habit performance varied significantly by individual and by habit. Performing a larger total number of habits was associated with stepwise improvements in both HbA1c% and TIR. Performance of a 7th (healthy eating) habit was associated with a further reduction in HbA1c%.
Conclusions: We have validated “the 6 habits” as a composite biomarker of engagement in T1D self‐management and identified a 7th habit that associates with glycemic control. Tracking these habits may assist providers with selecting interventions to improve glycemic control among youth.
Keywords: adherence, hemoglobin A1c, type 1 diabetes
Abstract #5 Creation of a diabetes data dock to integrate, improve, and analyze diverse data sources and facilitate continuous learning and improvement
Brent Lockee; Mitchell Barnes; Craig Vandervelden; Emily DeWit; Mark C. Clements
Children's Mercy Kansas City Kansas City, Missouri, USA
bclockee@cmh.edu
Background/Objective: Many diabetes centers within the T1D Exchange Quality Improvement Collaborative aspire to implement risk‐based approaches to care, yet they face challenges mapping electronic medical records (EMR) to the T1D Exchange data standard and integrating those data with diabetes self‐management device data and other sources of data to track and visualize population risk. We designed and constructed a diabetes clinic‐focused cloud data infrastructure (aka, the D‐Data Dock) characterized by highly reusable, general data processes that can support the diverse goals of diabetes research and clinical care.
Methods: The D‐Data Dock (Figure 1) was built using Azure cloud solutions and follows a data lakehouse architecture. The design allows us to scale and adjust processes to meet the volume, variety, velocity, and veracity demands of the data.
Results: Automating the data ingestion and mapping process has enabled our team to rapidly integrate novel risk biomarkers (e.g., 6 Habits of diabetes self‐management [JAMA Netw Open, 2021 Oct 1;4(10):e2131278] and CGM‐based risk biomarkers [Ped Diab, 2021 Nov 22(7):982–991]) into population health dashboards and statistical process control charts to transform clinical care delivery and research. We have used the D‐Data Dock to implement AI‐driven predictions of near‐term outcomes (e.g., hospitalization for ketoacidosis, 90‐day change in A1c, 30‐day change in time in range) and to pilot a remote patient monitoring program among high‐risk youth.
Conclusions: The D‐Data Dock is a replicable and scalable resource ready for widespread adoption by T1D Exchange member centers to improve population health management and accelerate clinical research.
Keywords: artificial intelligence, automatic, data processing, data quality, data sharing
Figure 1: The D‐Data Dock contains device data from the cloud platforms for six self‐management device manufacturers, device data aggregators for >3200 clinic patients, EHR and mapped T1D Exchange data for >5000 current and former patients, and supplemental data from other sources. Device data from all vendors are mapped to common schemas for glucose, insulin, and carbohydrates.
Abstract #6 Using quality improvement techniques to reduce delayed hospital follow‐up in youth with known type 1 diabetes admitted in diabetic ketoacidosis
Kelsee Halpin; Jude El Buri; Katie Noland; Regina S. Vergarabagby; Robin Kenyon; Cierra Reynolds; Andie Kaminsky; Erica Zarse; Jennifer W. Boyd
Children's Mercy Kansas City Kansas City, Missouri, USA
khalpin@cmh.edu
Background/Objective: Diabetic ketoacidosis (DKA) is a life‐threatening complication of type 1 diabetes (T1D). Guidelines recommend outpatient follow‐up within 1 month of hospital discharge for patients with diabetes hospitalized with hyperglycemia, including DKA. We identified that 63% of pediatric patients with known T1D admitted for DKA were being scheduled for outpatient follow‐up >28 days after their hospitalization. A quality improvement (QI) initiative was developed to reduce this percentage to less than 40%.
Methods: Using the QI tools of A3 Problem‐Solving Assessment Form and Process Flow Maps, interventions were developed, including (1) addition of 2 clinic appointments per week that were blocked solely for scheduling hospital follow‐ups; (2a) modification of inpatient charting to require the documenter to input the date of the follow‐up appointment, thus prompting scheduling before discharge; (2b) an alert added to documentation should the charted follow‐up date be >28 days to encourage timely follow‐up.
Results: Opening additional appointments during our first PDSA cycle reduced our follow‐up >28 days to 52% after 6 weeks. During our second PDSA cycle, which included our charting interventions, the follow‐up >28 days reduced further to 35% after 12 weeks. Continued assessment over 8 months showed persistent improvement with follow‐up >28 days averaging 41%.
Conclusions: Using QI tools to develop interventions to increase access and guide workflow, our team was successfully able to reduce the percentage of pediatric patients with delayed hospital follow‐up after admission for DKA.
Keywords: diabetic ketoacidosis, patient discharge, quality improvement, type 1 diabetes mellitus
Abstract #7 Utilization of a CGM‐based dashboard to identify at‐risk patients with type 1 diabetes (T1D)
Katie Noland1; Britaney Spartz1; Emily DeWit1, Mark Clements1; Rachel Dixon1; Jaimie Contreras1; Gayla Kutzli1; Andie Kaminsky1; Katelyn Evans1; Jude El Buri1
1Children's Mercy Kansas City Kansas City, Missouri, USA
kenoland@cmh.edu
Background/Objective: A continuous glucose monitor (CGM)‐based population health dashboard, created by researchers at Stanford University, was adapted and adopted by Children's Mercy. The dashboard flags patients with T1D meeting clinician‐defined risk criteria: extreme lows >2%, no alerts, lows >4%, >15% drop‐in time‐in‐range (TIR), TIR <65%, extreme highs >10%, >15% drop in wear time, insufficient data, extreme highs >3%. It enables clinicians to identify patients in high‐risk categories for additional support between standard‐of‐care visits.
Methods: We implemented the dashboard using Power BI and conducted four PDSA cycles targeting biomarker‐based risk groups. Families received a one‐time phone call or were scheduled for a series of problem‐solving calls.
Results: Figure 1 shows the results of telephone outreach across four PDSA cycles. 33% of patients were flagged based on out‐of‐date CGM data, had already transitioned to adult care, or had time in range >80% despite meeting other risk biomarkers. These issues affecting efficient workflow and patient acceptability were addressed after Cycle 2. Subsequently, 0% of patients who were reviewed in the dashboard were affected by these issues. Overall, families were more engaged in making insulin changes at the point of contact than when scheduling a series of future remote contacts.
Conclusions: A CGM‐based risk dashboard may help clinicians identify patients who would benefit from proactive outreach between in‐clinic visits, but families may be difficult to reach by phone, and some families may not perceive that they need help. Future work should seek to overcome these barriers.
Keywords: CGM, diabetes, population health management
Figure 1: Proportion of individuals identified as high‐risk in the CGM risk dashboard who (A) were reached by telephone, (B) declined assistance, and (C) either scheduled future remote patient monitoring visits (QI staff phone calls) or accepted immediate assistance (Clinic staff phone calls).
Abstract #8 Addressing social determinants of health in an ambulatory pediatric diabetes clinic; examining data by race and ethnicity
Emily DeWit, Brent Lockee, Mitchell Barnes, Katelyn Evans, Mark Clements, Kelsee Halpin, Shilpi Relan
Children's Mercy Kansas City Kansas City, Missouri, USA
eldewit@cmh.edu
Background/Objective: The American Diabetes Association has recommended screening for social determinants of health (SDOH) and addressing social barriers to health for all individuals with diabetes. Our aim was to implement a SDOH screening tool in a pediatric type 1 diabetes (T1D) clinic, then analyze the completion and positivity rates by race and ethnicity.
Methods: In September 2021, we implemented a SDOH screening survey in clinic intake forms. We completed 10 Plan Do Study Act cycles by July 1, 2022. Cycles tested delivering a resource list to those with positive screens, providing a link to a web‐based platform with comprehensive resources, and providing personalized guidance to connect to vetted community resources. We also modified language to give families a rationale for taking the survey and to articulate that guidance service was free.
Results: The survey was completed by 3370 out of 4071 families attending clinic appointments from September 1, 2021 to July 1, 2022. Overall, 4.42% of screens were positive. Whites completed 78.9% of surveys yet made up 43.62% of positive screens. In contrast, Black individuals completed 7.83% of surveys yet accounted for 28.19% of positive screens. Hispanic/Latinos completed 9.17% of surveys while accounting for 14.77% of positive screens.
Conclusions: Black and Hispanic patients are disproportionately affected by SDOH barriers. This disparity illustrates the need for screening and addressing barriers to SDOH. Screening for SDOH should drive the development of cost‐effective, culturally customized programs to support diabetes care and promote health equity.
Keywords: racial disparities, social determinants of health, type 1 diabetes
Abstract #9 Continuing improvement of health equity: Use of continuous glucose monitoring technology among youth with type 1 diabetes
Kajal Gandhi; Kathryn Obrynba; Justin Indyk; Don Buckingham; Manmohan Kamboj
Kajal Gandhi, DO, MPH – Attending Physician, Nationwide Children's Hospital Columbus, Ohio, USA
kajal.gandhi@nationwidechildrens.org
Background/Objective: The use of continuous glucose monitoring (CGM) improves glycemic outcomes and quality of life among youth with type 1 diabetes (T1D). Furthermore, health disparity research reveals that those belonging to racial/ethnic minorities have consistently lower CGM utilization. This quality improvement project aims to continue the initiative to reduce the disparity among Black/White pediatric T1D populations' uptake of CGM technology at Nationwide Children's Hospital.
Methods: Our multidisciplinary quality improvement clinical team applied Institute of Healthcare Improvement’s Model of Improvement using statistical process, using statistical process control charts to measure change effectiveness. The total patient clinic census was evaluated for 22 months for indicators of both new CGM uptake and continued use among T1D patients with a duration of T1D ≥ 12 months. Where eligible patients had not yet adopted CGM, education and uptake assistance were provided with sensitivity for racial bias. CGM uptake disparity ratio, defined as the ratio of Black/White population use of CGM per 100 clinic visits, was calculated, with the equity goal ratio being 1:1.
Results: CGM use was compared across T1D patient populations. Among White and Black youth T1D populations, the equity disparity ratio improved from a baseline rate per 100 visits from ratios of 0.71:1.0, to 0.87:1.0 over the study period.
Conclusion: Data showed: (1) overall increase of CGM technology use across all T1D populations; (2) approximate 50% equity uptake improvement between the Black and White new users, further closing the disparity gap in access to technology. Future work includes improving the disparity among the Black/White T1D population and expanding to other ethnic minority populations with T1D to improve diabetes outcomes.
Keywords: continuous glucose monitoring, health equity, type 1 diabetes mellitus
Figure 1.
Abstract #10 Artificial intelligence decision support enhances engagement and integration with home diabetes care
Gajanthan Muthuvel; Siobhan Tellez; Katherine Bowers, Patrick Brady; Nancy Daraiseh; Jane Khoury; Emily Smith; Sarah Corathers
Cincinnati Children's Hospital Medical Center, University of Cincinnati Department of Pediatrics Cincinnati, OH, USA.
gajanthan.muthuvel@cchmc.org
Background/Objective: Insulin adjustments outside of the clinic are important for improving glycemic outcomes in patients with type 1 diabetes (T1D). However, interval changes are inconsistently performed, in part due to the complexity and volume of data generated by diabetes devices hindering self‐interpretation and efficient provider interactions. We aimed to examine the feasibility, acceptability, and efficacy of an enhanced care intervention (ECI) that uses artificial intelligence (AI) decision support tool, DreaMed Advisor Pro®, to augment insulin dose adjustment recommendations delivered between visits for youth with T1D.
Methods: Baseline, 3 and 6‐month visits were conducted with monthly reminders to upload diabetes devices for inter‐visit AI‐guided insulin recommendations. Eligibility criteria: T1D for ≥6 months, use of Omnipod pump and Dexcom continuous glucose monitor, age 7–24 years, and baseline HbA1c or glucose management indicator 7%–13%. Outcomes for pre‐post intervention analysis: HbA1c, time in range (TIR), % hypoglycemia episodes <70 mg/dl, and validated survey instruments assessing acceptability, appropriateness, and feasibility.
Results: No differences were detected for HbA1c (p = 0.94), TIR (p = 0.49), or hypoglycemia (p = 0.09). However, feedback was largely positive regarding the acceptability, feasibility, and appropriateness of the ECI (Table 1). Challenges at times were seen with home upload compatibility and data flow between diabetes platforms, limiting the use of AI decision support.
Conclusions: Family feedback suggests enthusiasm in interventions designed to support home diabetes care between visits. Decreasing the onus required for home uploads to maintain participant investment and reducing friction in communication between unique software platforms may aid in optimizing AI‐guided decision support to have a greater impact on glycemic trends.
Keywords: artificial intelligence, diabetes technology
Table 1: Glycemic outcomes were compared over 6 months using a general linear mixed model for pre‐post intervention analysis, as well as assessment of acceptability, appropriateness, and feasibility with validated surveys.
| Variable | Baseline (n = 96) | 3‐month visit (n = 92) | 6‐month visit (n = 75) |
|---|---|---|---|
| HbA1c (%) [mean ± std (range)] | 7.79 ± 0.87 (6.1, 10.6) [n = 95] | 7.84 ± 1.05 (5.9, 11.1) [n = 90] | 7.81 ± 0.87 (6.0, 10.2) [n = 72] |
| Glucose management index (%) [mean ± std (range)] | 7.96 ± 0.74 (6.7, 10.2) [n = 88] | 7.95 ± 0.82 (6.6, 10.9) [n = 83] | 8.03 ± 0.66 (6.7, 10.0) [n = 62] |
| Time in range (%) [mean ± std (range)] | 45.4 ± 14.3 (5.5, 72.1) [n = 96] | 46.3 ± 15.5 (2.1, 87.6) [n = 90] | 43.6 ± 12.7 (13.0, 74.60 [n = 74] |
| Hypoglycemia (%) [mean ± std (range)] | 1.96 ± 2.24 (0, 11.4) [n = 96] | 2.87 ± 6.18 (0, 57.2) [n = 90] | 2.16 ± 2.10 (0, 9.7) [n = 73] |
| General Patient/Family Feedback | ‐ | n = 58 | n = 49 |
| Were the recommendations helpful? | 84.5% (49) agree | 83.7% (41) agree | |
| Acceptability | ‐ | n = 84 | n = 73 |
| The ECI meets my approval | 84.5% (71) agree | 91.8% (67) agree | |
| The ECI is appealing to me | 88.1% (74) agree | 95.9% (70) agree | |
| I like the ECI | 79.8% (67) agree | 90.4% (66) agree | |
| I welcome the ECI | 90.5% (76) agree | 97.3% (71) agree | |
| Appropriateness | ‐ | n = 84 | n = 73 |
| The ECI seems fitting | 85.7% (72) agree | 95.9% (70) agree | |
| The ECI seems suitable | 88.1% (74) agree | 95.9% (70) agree | |
| The ECI seems applicable | 89.3% (75) agree | 93.2% (68) agree | |
| The ECI seems like a good match | 84.3% (70/83) agree | 90.4% (66) agree | |
| Feasibility | ‐ | n = 84 | n = 73 |
| The ECI seems implementable | 85.7% (72) agree | 93.2% (68) agree | |
| The ECI seems possible | 89.3% (75) agree | 95.9% (70) agree | |
| The ECI seems doable | 89.3% (75) agree | 94.5% (69) agree | |
| The ECI seems easy to use | 83.3% (70) agree | 89.0% (65) agree |
Abbreviation: ECI, enhanced care intervention.
Abstract #11 Quality improvement in action: Addressing continuity of care in pediatric endocrinology fellowship
Jonathan D. Tatum; Alison Murray; Samantha H. Roberge; Varsha M. Thomas; David Baute; Andrew Lavik; Amy Grant; Sarah D. Corathers
Cincinnati Children's Hospital Pediatric Endocrinology Cincinnati, Ohio, USA
jonathan.tatum@cchmc.org
Background/Objective: Continuity of care is important in medical training and patient experience, particularly with chronic conditions like diabetes. The aim of this quality improvement (QI) project was to increase the percentage of patients seen in pediatric endocrinology fellows' continuity clinic that are subsequently scheduled with the same provider from 38% to 60% over 4 months.
Methods: A multi‐disciplinary team of endocrine fellows, scheduling, and clinic staff studied existing processes of scheduling and developed a key driver diagram reflecting our theory for improvement. We conducted iterative Plan‐Do‐Study‐Act cycles and tracked outcomes on a run chart. Continuity was defined as a patient scheduling with the same fellow provider on a subsequent visit by 2 weeks after the current visit. Graduating fellows' data were excluded during the intervention period because they did not have the opportunity for future visits to be scheduled.
Results: Continuity improved from a baseline 38% to 74% over 4 months, exceeding the target. The most impactful interventions included fellows filling out a previously under‐used “Follow‐up” section on the electronic medical record that is seen by the scheduling center, visual reminders for families to schedule their next appointment during the current visit, and huddles with clinic nurses to confirm the desired follow‐up.
Conclusions: Fellow continuity was improved using PDSA cycles and applying QI methodology, exceeding the stated goal. As a result, 4 of 5 patients seen by a fellow were scheduled with that same fellow in the future as compared to the previous 2 of 5 visits.
Keywords: continuity of care; fellowship; follow‐up; quality improvement
Figure 1.
Abstract #12 In sync with diabetes: Increasing remote upload of insulin pump data among children with type 1 diabetes
Andrew R. Lavik; Sonia Priscila Rodas Márquez; Amy Grant; Monica Gumz; Tammy DiMuzio; Sarah D. Corathers; Allison Deisinger; Karishma Tilton; Kimberly Keppler; and Gail Patten
Department of Pediatrics, Division of Endocrinology; Cincinnati Children's Hospital Medical Center Cincinnati, Ohio, USA
andrew.lavik@cchmc.org
Background/Objective: Insulin pump data upload is crucial for pediatric type 1 diabetes (T1D) care delivery, historically performed at in‐person visits. However, data upload by staff requires significant time, extends visit length, and cannot be performed at telemedicine visits. Methods for remote data upload exist but are underutilized. We, therefore, designed a quality improvement (QI) project with a SMART Aim to increase the percentage of insulin pumps uploaded prior to T1D clinic visits at our institution from 15% to 30% by June 30, 2022.
Methods: Our team of clinicians, educators, parents, and QI specialists employed QI methods to understand the baseline state of pump upload and develop a key driver diagram reflecting our theory for improvement. We conducted iterative Plan‐Do‐Study‐Act cycles targeting key drivers and tracked outcomes on a run chart.
Results: We increased the percentage of insulin pumps uploaded prior to visiting from 15% to 33% (Figure 1). The most effective interventions included discussing the goal of pre‐visit pump upload with families, emailing upload instructions to families 1 week before the visit, and syncing all new pumps to our professional cloud account during pump start.
Conclusions: The use of QI methods increased the percentage of pediatric T1D patients whose insulin pump data was uploaded prior to clinic visits by 18%, corresponding to an additional 45 patients monthly with data uploaded pre‐visit. This process not only streamlines visits, but also allows remote data review by providers and families, which is increasingly important in this digital age.
Keywords: insulin infusion systems, quality improvement, type 1 diabetes
Figure 1. Percentage of insulin pumps uploaded prior to T1D visit.
Abstract #13 Equity in diabetes care & transformation: Implementation & spread of social determinants of health screening in diabetes clinic
Nana‐Hawa Yayah Jones; Sarah Corathers; Amy Grant; Jennifer Kelly; Molly Williams; Kyle Kaplan; Kelsey Hart; Mona Mansour
Department of Pediatrics, Division of Endocrinology; Cincinnati Children's Hospital Medical Center Cincinnati, Ohio, USA
nana.jones@cchmc.org
Background/Objective: Morbidity and mortality in type 1 diabetes (T1D) are grossly marred by key disparities and equity gaps compounded by social determinants of health (SDH): non‐medical factors that influence health outcomes. This project aimed to increase the percentage of T1D patients screened for SDH in diabetes clinics from 0% to 90% by June 30, 2022. We additionally sought to co‐develop a patient‐centered, collaborative SDH tool to spread hospital‐wide in all healthcare conditions.
Methods: Using quality improvement (QI) methodology, a multidisciplinary diabetes team identified an SDH tool. By using plan‐do‐study‐act (PDSA) cycles, the tool matriculated into the standard of care. Once adapted from verbal to paper to automatic firing on electronic tablets, learnings were translated into a hospital‐wide core SDH screener intended for use in all ambulatory practices of the hospital, spreading beyond the T1D clinic.
Results: In total, 4502 screens have been completed in 9402 encounters (48%). With automation, the percentage of patients with complete SDH screens has risen from 0% to 88% (Figure 1). Up to 5% of the T1D population identified at least 1 SDH per visit. Cincinnati Children's Hospital cares for 2227 active T1D patients, of which 410 are publicly insured (18%). Of patients screening positive, 23% are publicly insured.
Conclusions: Implementation of an SDH screen in a diabetes clinic can be standardized and effective in identifying barriers to the healthcare needs of patients with diabetes. SDH screening in the T1D population can be readily adapted to screen other pediatric chronic disease populations in large tertiary care centers, magnifying its impact on healthcare outcomes.
Keywords: diabetes, equity, pediatrics, quality improvement, social determinants of health
Figure 1. Percent of patients with social determinants of health screens completed during a clinic visit from the start of the project to present date.
Abstract #14 Increasing patient engagement through the use of online diabetes questionnaire
Mouhammad Alwazeer, Susan Hsieh, Christin Morell, Stephanie Ogburn, Candice Williams
Cook Children's Medical Center Fort Worth, Texas, USA
mouhammad.alwazeer@cookchildrens.org
Background/Objective: The objective of our project is to improve the utilization of the online diabetes questionnaire (ODQ) by patients with type 1 diabetes prior to their clinic appointment. We aimed to increase the utilization of ODQ prior to the appointment by 10%.
Methods: Patients sign on to the EHR system via a portal to pre‐fill a health online diabetes questionnaire (ODQ) with regards to insulin dosing, carb counting, etc. The following interventions were tested:
Intervention 1: Send patient portal messages 3–7 days prior to appointment to remind patient/family to fill out the questionnaire
Intervention 2: The clinic secretary verbally reminded patients/families to fill out ODQ using smartphones after check‐in to eliminate notification fatigue
Intervention 3: The medical assistant (MA) reminded patients/families to fill out the ODQ while in the patient's room
Results: Following a series of rapid PDSA cycles, the percentage of patients utilizing the ODQ increased from a baseline of 37% to 58%. This represents an increase of 21% over 12 months.
Conclusions: Completing the ODQ prior to the appointment helps improve the quality of visits as it allows for more time with the provider. Surprisingly, online reminders to complete ODQ did not improve participation in ODQ prior to the visit. This resulted in notification fatigue. The study showed the most effective method to increase participation was onsite reminders via the MA, as it helped utilize patients waiting time efficiently. This will be expanded to all endocrine clinics at Cook Children's.
Keywords: EHR system, questionnaire
Abstract #15 Implementation of SDOH screening in patients with diabetes: Cook children's endocrinology clinic
Christin Morell, Susan Hsieh, Mouhammad Alwazeer, Stephanie Ogburn, Candice Williams, CDCES Cook Children's Medical Center
Fort Worth, TX, USA
christin.morell@cookchildrens.org
Background/Objective: Social determinants of health (SDOH) and conditions in which children are born, grow, and live affect children's health and contribute to health disparities. Cook Children's Endocrinology clinic did not have a formal process for screening for SDOH in patients with type 1 diabetes. The objective of this project was to implement SDOH screening. We aimed to screen 50% of patients over 6 months.
Methods: We created a multidisciplinary team that decided on utilizing the Hunger Vital Sign™. Identified patients were given a paper questionnaire with two Hunger Vital Sign™ questions at check‐in. The following interventions were tested in rapid PDSA cycles:
Provided survey both in English and Spanish, and screening is completed based on patient/caregiver's preferred language
Included statement to assure caregiver that responses would be confidential
Connected positive screens to available resources
Results: Before the implementation of the survey, there was no screening data regarding food insecurity. Following interventions, the survey completion rate increased from 0% to 85% over 6 months.
Conclusions: Implementing a formal screening process using a validated tool allowed the discovery of patients with food insecurities. Offering SDOH screening in both English and Spanish, and assuring patients/caregivers that responses are confidential helps increase the completion rate. Referral of positive screens to local food resources supported patients in managing their diabetes better. This result supports the need for expansion of screening to all patients with diabetes in the Cook's Children satellite clinic location.
Keywords: food insecurity, hunger vital sign, SDOH
Figure 1
Abstract #16 Improving access to continuous glucose monitors for patients with T1D in a safety net hospital clinic
Alisha Virani2; David C. Ziemer1; Georgia M. Davis1; Francisco J. Pasquel1; Kristi Quairoli2; J. Sonya Haw1
1Emory University School of Medicine Division of Endocrinology, 2Grady Health System, Diabetes Center Atlanta, Georgia 30303 USA
avirani@gmh.edu
Background: The use of continuous glucose monitors (CGM) is low among adult type 1 diabetes (T1D) populations from historically disadvantaged backgrounds. We aimed to increase CGM use by improving the process flow of CGM paperwork.
Methods: We examined the usage of CGM before and after the implementation of PDSA cycles aimed at improving technology access among patients receiving care at the Grady Diabetes Center. Sequential processes were implemented after identifying barriers to CGM use. Barriers included the morass of paperwork for approval and distribution, flawed tracking systems for paper documents, required insurer questionnaires, and unexplained denials. Implementation steps included assigning a single provider streamlining paperwork into one location, recruiting our social worker to manage paperwork, and migrating into Parachute Health, an electronic application and tracking service for durable medical equipment. We estimated the proportion of patients using CGM before and after these PDSA cycles.
Results: 285 adults with T1D were included during the implementation period. Most patients were African American (89%), 53% female, and 73% with ages between 18–50 years. Most patients had federal or no health insurance coverage (46% Medicaid, 19% Medicare, 18% uninsured, and 16% commercial). Site data mapped in the T1D Exchange QI Portal showed an increase in CGM use from 12% to 57% from July 2020 to June 2022.
Conclusions: Improving the process flow of CGM paperwork led to increased CGM prescriptions among patients with T1D receiving care at an inner‐city safety‐net hospital. Future process improvement interventions will investigate the efficacy of translating these prescriptions into ongoing patient CGM use.
Keywords: blood glucose self‐monitoring, health equity, medical device, safety‐net hospitals, type 1 diabetes
Abstract #17 The diabetes, data & devices (D3) education program: Motivating patient‐driven review of diabetes data and insulin changes
Ashley Garrity; Inas Thomas; Jacqueline Fisher; Emily Hirschfeld; Joyce M. Lee
C.S. Mott Children's Hospital, University of Michigan Ann Arbor, Michigan, USA
ashleyna@med.umich.edu
Background/Objective: Despite increasing use of continuous glucose monitoring (CGM) and pump technology, patient‐driven review of data and changing of insulin doses (Habits 5 and 6) are rare, despite their association with improved glycemic outcomes. This project aimed to provide patients and families with training in how to interpret and take action with their data.
Methods: Through a literature review and qualitative work with diabetes providers, an interdisciplinary diabetes team developed the Diabetes, Data & Devices (D3) curriculum. The D3 program promotes regular retrospective data review and trains patients and families to understand key glycemic metrics, identify patterns of hypoglycemia or hyperglycemia in their data, and proactively make changes in behavior and/or insulin dosing. Pre‐ and post‐class glycemic outcomes, diabetes management self‐efficacy, and diabetes data literacy were evaluated.
Results: To date, 151 pediatric diabetes patients and their families have attended a D3 class. We excluded new‐onset patients within 3 months of diagnosis (n = 113). Post a minimum of 150 days after attending class, patients had an average reduction in hemoglobin A1c (HbA1c) of 0.37% with variation by age: 0–5 yrs (−0.82%); 6–11 yrs (−0.20%); 12+ years (−0.35). The total self‐efficacy score increased by 1.2 points and the data literacy score increased by 0.3 points.
Conclusions: The D3 program educates and empowers families to review and take action with their personal diabetes data, leading to increased knowledge and confidence and improvements in glycemia.
Keywords: data literacy, diabetes self‐management, retrospective data review
Abstract #18 2021 T1D Exchange QI clinic practice survey: Clinic staffing and structure for adult and pediatric diabetes clinics
Mary Pat Gallagher, Emma Ospelt, Nicole Rioles, Ruth S. Weinstock, Devin Steenkamp, Grazia Aleppo, Elizabeth A. Mann, Sonya Haw, Jenise Wong, Osagie Ebekozien
The Hassenfeld Children's Hospital, Pediatric Diabetes Center at NYU Langone New York, New York, USA
marypat.gallagher@nyulangone.org
Background/Objective: The T1D Exchange Quality Improvement Collaborative (T1DX‐QI) collected data on staffing ratios from participating clinics to describe the current state of adult and pediatric diabetes practices. This data will assist in creating benchmarks to optimize care for pediatric and adult populations with T1D.
Methods: Surveys were collected from October 5 to November 19, 2021, from 34 T1DX‐QI sites (25 pediatric, nine adult). Full‐time equivalents (FTEs) of different disciplines at each diabetes clinic were reported. Data were cleaned and analyzed in R studio. FTEs of diabetes team members from different disciplines per 1000 patients were compared in adult versus pediatric sites using t‐tests. Differences in support staff, weekend, and evening staff between pediatric and adult sites were also evaluated.
Results: When analyzed per 1000 patients, the FTEs were greater in pediatric (n = 25) compared to adult sites (n = 9) in every discipline. In adult clinics, the lowest staffing ratios were in social work (0.09 FTE/1000 patients) and psychology (0.06 FTE/1000 patients), versus 0.9 and 0.4 FTE, respectively, in pediatric sites. The ratio of most support staff did not differ between pediatric and adult sites.
Conclusions: These data are consistent with those collected by T1DX‐QI in 2015, demonstrating persistent staffing disparities between adult and pediatric clinics. Inequities are particularly notable in behavioral health support. Sixteen percent of pediatric sites reported nurse practitioner participation in evening/weekend coverage, which may reflect the increased need for fellowship applicants in pediatric endocrinology. Only half of the available fellowship positions were filled through the NRMP Pediatric Specialties Match in 2021.
Keywords: benchmarking, quality improvement, type 1 diabetes mellitus, workforce
Table 1. Total FTE of diabetes team members among pediatric and adult clinics in the T1DX‐QI
| Total FTE of diabetes team members | Overall clinic average FTE N = 34 | Pediatric clinic average FTE N = 25 | Adult clinic average FTE N = 9 | Pediatric clinic FTE average per 1000 patients [SD] | Adult clinic FTE average per 1000 patients [SD] | Pediatric vs. adult FTE average per 1000 patients p‐value† |
|---|---|---|---|---|---|---|
| MD/DO | 7.87 | 8.0 | 7.6 | 3.6 [2.4] | 1.6 [1.5] | 0.001 |
| NP/PA | 3.31 | 3.4 | 3.1 | 1.5 [0.8] | 0.7 [0.6] | 0.003 |
| Social Worker | 1.63 | 2.0 | 0.5 | 0.9 [0.6] | 0.09 [0.2] | <0.001 |
| RN | 4.59 | 5.3 | 2.7 | 2.3 [1.7] | 0.5 [0.8] | <0.001 |
| CDCES | 6.32 | 7.3 | 3.8 | 3.3 [1.7] | 0.8 [1.1] | <0.001 |
| Psychology | 0.7 | 0.8 | 0.3 | 0.4 [0.4] | 0.06 [0.2] | 0.003 |
Abstract #19 Improving SDOH screening: the HCH pediatric diabetes center at NYU
Jeniece Ilkowitz, Juana Gonzalez, Kathryn Licciardello, Mary Pat Gallagher
The Hassenfeld Children's Hospital, Pediatric Diabetes Center, at NYU Langone New York, New York, USA
jeniece.ilkowitz@nyulangone.org
Background/Objective: Social Determinants of Health (SDOH) impact diabetes management and are related to the well‐documented disparities in diabetes outcomes. SDOH screening can identify families in need of resources. Given that food and housing insecurity and other SDOH impact diabetes management, screening for SDOH is an essential aspect of diabetes care. This quality improvement (QI) project was initiated at the HCH Pediatric Diabetes Center (PDC) to increase annual documentation of SDOH screening by 50% after 1 year.
Methods: To increase SDOH screening, we added questions to the PDC patient intake form. Intake form distribution began in January 2021. We used a series of Plan‐Do‐Study‐Act (PDSA) cycles to help increase the completion rate. We initially administered the screen on paper at check‐in, changed it to email in advance of the appointment, then to a MyChart questionnaire in advance of the appointment, and then via Welcome Tablets at check‐in.
Results: Following a series of PDSA cycles to improve intake form completion, including documentation of SDOH screening, completion rates increased from 2% to 70% in 12 months.
Conclusions: Using an intake form with multiple PDSA cycles to determine the most effective form of delivery increased documentation of SDOH screening over 12 months. Using a Welcome Tablet at check‐in was the most effective in capturing SDOH screening at our PDC.
Keywords: diabetes mellitus, type 1, pediatrics, quality improvement, social vulnerability, surveys and questionnaires
Abstract #20 Implementation of a wellness program for people with HbA1c >9%
Jeniece Ilkowitz, Vanessa Wissing, Chris Lally, Mary Pat Gallagher
The Hassenfeld Children's Hospital, Pediatric Diabetes Center at NYU Langone New York, New York, USA
jeniece.ilkowitz@nyulangone.org
Background/Objective: People with elevated HbA1c levels are at higher risk for diabetes complications. The HCH Pediatric Diabetes Center (PDC) created the Wellness Program (WP) to provide additional support to people with HbA1c levels >9%. Our goal was to decrease HbA1c levels to <9% for ≥20% of participants within 6 months.
Methods: This quality improvement (QI) initiative began in August 2021. We assigned people with HbA1c >9% to a CDCES who confirmed/scheduled a WP appointment. During visits, we assessed barriers and reviewed the benefits of lowering HbA1c levels. Through shared decision‐making, we identified and tested individualized interventions, including weekly remote CDCES check‐ins, a six‐week in‐person NP/MD follow‐up, the ability to text/call a dedicated WP cellphone, support to increase diabetes technology use, and behavioral health/psychosocial supports. We reviewed outcomes at 6 months.
Results: Of our population (n = 464), 82 (17.6%) had HbA1c levels >9% (mean = 10.6 ± 1.5%, median = 10.1% [9.2–16]). The median age was 14 years in both groups. Of the WP population, 70.2% self‐identified as Non‐White versus 31.4% of the total population. At 6 months, 32 (39%) had an HbA1c <9% with a mean post‐program HbA1c level of 8.2 ± 0.6% (median = 8.4% [6.5–8.9]). Interventions tested included: initiation of Control‐IQ (n = 6), diabetes education/management (n = 5), behavioral health intervention/additional psychosocial care in place (n = 4), increased communication (n = 6), and other/unknown (n = 6).
Conclusions: The WP QI effort decreased the HbA1c of 39% of those enrolled to <9% within 6 months. Ongoing initiatives include further investigation into observed racial disparities, interviews at WP exit, and weekly PDC team review of WP participants' progress.
Keywords: diabetes mellitus, pediatrics, quality improvement
Abstract #21 Review of common errors during EMR data mapping and transformation to the T1D Exchange QI data specification
Anton Wirsch, Ann Mungmode, Saketh Rompicherla, Emma Ospelt, Nudrat Noor, Joyce Lee, Marina Basina, Jesse Cases, Osagie Ebekozien
T1D Exchange, Boston, Massachusetts, USA
awirsch@t1dexchange.org
Background: The T1D Exchange Quality Improvement Collaborative (T1DX‐QI) is a network of 49 endocrinology data‐sharing centers collaborating with the goal of improving care for people with type 1 diabetes. T1DX‐QI is a Learning Health System where participating centers transfer de‐identified electronic medical records (EMR) after data mapping, transformation, and validation for population health improvement.
T1DX‐QI centers map existing data fields in their respective EMR to a unified T1DX‐QI data specification (T1DX‐DS). T1DX‐DS contains 120 variables across seven files, including the Patients, Providers, Encounters, Observations, Conditions, Medications, and Diabetes Files. Errors in the mapping and transformation process lead to a significant delay in the onboarding process.
Method: The authors reviewed the initial files shared from 11 centers and feedback documentation from the T1DX‐QI team following the initial data mapping and transformation process. These errors were categorized by T1DX‐DS file and common error types.
Results: All the reviewed centers had initial errors of Value Case, Value Coded, and Value Incorrect (Figure 1 left). Seventy‐three percent of the centers had errors in the diabetes file. All clinics experienced mapping errors in the Encounter, Observation, and Patient files (Figure 1 right).
Conclusion: Our review identified common data mapping and transformation errors in the T1DX‐DS. The errors could be due to several factors, which include the clarity of the data specifications, local clinic capacity, nature of the data, and technical challenges. T1DX‐QI data team is testing multiple improvement interventions to reduce these common errors in the future and further streamline the data mapping process.
Figure 1
Abstract #22 User value: comparisons between data‐mapped and non‐mapped T1D Exchange quality improvement portal users
Ann Mungmode; Roberto Izquierdo; Don Buckingham; Julie Samuels; Anna Neyman; Osagie Ebekozien
T1D Exchange Boston, Massachusetts, USA
amungmode@t1dexchange.org
Background/Objective: The T1D Exchange Quality Improvement Collaborative (T1DX‐QI) is a network of over 45 endocrinology centers aiming to improve care for people with type 1 diabetes. Participating centers have access to the QI Portal, an EMR‐based data platform with four main features: (1) Dashboard (overview of center outcomes), (2) Compare (center‐to‐center benchmarking), (3) Report (center‐specific outcome charting), and (4) Library (QI case studies, changes packages, and tools). Data‐mapped (DM) users have access to all four feature tabs. Non‐mapped (NM) users are not yet providing monthly center data to T1DX‐QI and have access to a limited version of the QI Portal (Compare and Library tabs). The goal of this study was to compare QI Portal use between DM and NM users.
Methods: We included returning users (excluding the first‐time user logins) as a measure of value‐based interaction with the QI Portal. The number of returning users per month and average time spent were tracked in Google Analytics and analyzed visually in Figure 1.
Results: On average, DM users spend more time on applicable QI Portal tabs than NM users. A higher percentage of DM users returned to the Dashboard and Report tabs, whereas a higher rate of NM users returned to the Compare and Library tabs.
Conclusions: These results indicate that there is value in endocrinology centers having access to chart and share real‐world data outcomes. Users spend the most time manipulating center‐specific data in the Reports tab, followed by the benchmarking Compare tab.
Keywords: data, electronic medical record, product development, quality improvement, type 1 diabetes
Figure 1. Percentage of users and time spent by QI Portal Tab
Abstract #23 Building capacity for quality improvement
Ori Odugbesan; Ann Mungmode; Nicole Rioles; Holly Hardison; Osagie Ebekozien
T1D Exchange, QI, and Population Health, Boston, Massachusetts, USA
oodugbesan@t1dexchange.org
Background/Objective: The T1D Exchange Quality Improvement Collaborative (T1DX‐QI) is a type 1 diabetes quality improvement (QI) collaborative that consists of 49 pediatric and adult endocrinology clinics. Each participating site identifies a QI Coordinator to lead activities and support the use of QI tools. T1DX‐QI has provided resources to support QI Coordinators based on their QI knowledge and expertise. T1DX‐QI combined four broad areas to support QI capacity building using coaching, a QI committee, online learning, and data sharing. We aimed to build the QI capacity of Coordinators, make QI an integral part of clinic daily operations and improve collaborative performance on QI metrics.
Methods: T1DX designed a survey using a combination of Likert‐type, multiple‐choice, and open‐ended questions to understand the perspective of QI Coordinators on the resources provided and how duration and participation in the T1DX‐QI have contributed to capacity building. The survey was administered to 40 QI Coordinators using Qualtrics.
Results: Analysis was based on 32 completed surveys. The majority (79%) were pediatric site Coordinators, and 21% were adult site Coordinators. Nearly all of the respondents joined the collaborative within the last 3 years (95%), and 96% agreed that participating in the T1DX‐QI increased their QI knowledge. All (100%) respondents reported feeling empowered to use QI tools.
Conclusion: Providing QI resources and tailored support are key elements enabling a clinical team's commitment to QI. The T1DX‐QI capacity‐building approach is feasible; continued expansion of resources will improve the quality of patient care. Opportunities to engage leadership and funding should be further explored to support capacity building.
Figure 1.
Abstract #24 Increasing access to continuous glucose monitor: addressing health disparities in type 1 diabetes
Christy Byer‐Mendoza; Ori Odugbesan; Kim McNamara; Andrea Huber; Erin Dale, Erin Carpenter, Drisana Moss, Karen Anaya, Yashia Saenz; Carla Demeterco‐Berggren
Rady Children's Hospital, San Diego, California, USA
cbyer-mendoza@rchsd.org
Background/Objective: Continuous glucose monitor (CGM) use is associated with improved outcomes in type 1 diabetes (T1D), however, there remain significant disparities in the use of diabetes technologies. This study aimed to increase the percentage of children and adolescents with T1D wearing a CGM from a baseline of 64% to 80% in 12 months.
Methods: As a participating clinic of the T1D Exchange Quality Improvement Collaborative, we obtained monthly aggregate data over the intervention period using the electronic health record. A multidisciplinary team was formed to define the existing process. With patient, provider, and staff feedback, we identified key change concepts to reduce disparities and increase access to CGM. These were tested in a series of PDSA cycles. Interventions included: pharmacy technician and diabetes team assisting patients with documentation requirements, advocating for change in Medicaid's requirements, care navigators ‘outreach with a whole child model of care approach facilitating access to care, child life specialist telehealth visits to improve consistent CGM use, workflow implementation to initiate CGM at the inpatient setting during diagnosis.
Results: As of February 2022, the percentage of patients wearing a CGM increased by 23% from the February 2021 baseline of 64% to 87% (goal: 80%), reaching 88% in April 2022. Figure 1.
Conclusions: Family and patient‐centered interventions, support for social determinants of health that may impact diabetes technology access, staff training, local and state‐level advocacy, and efficient workflow substantially increased CGM use among all children and adolescents with T1D.
Keywords: CGM, health disparities, type 1 diabetes
Figure 1. CGM rates
Abstract #25 Increasing frequency of clinic visits among medicaid insured children and adolescents with type 1 diabetes
Carla Demeterco‐Berggren; Giana Reuter; Kim McNamara; Jodi Peterson; Yashia Saenz; Drisana Moss, Karen Anaya, Rhonda Sparr Perkins; Keri Carstairs; Michael Gottschalk
Rady Children's Hospital, San Diego, University of California, San Diego, San Diego, California, USA
cdemeterco@rchsd.org
Background/Objective: The relationship between more frequent clinic visits and improved HbA1c levels in people with type 1 diabetes (T1D) has been reported. The ADA recommends quarterly follow‐up for routine diabetes care, but this is not always attained. This quality improvement (QI) study aimed to increase the percentage of Medicaid‐insured children who attended at least four diabetes clinic visits per year (in person, telemedicine, or telephone) from a baseline of 53% to 75% in 12 months.
Methods: A diabetes clinic dashboard within our electronic health record (EHR) system allowed our team to identify children due for a clinic visit without a future appointment scheduled from a subset of the Medicaid‐insured cohort (n = 100). A multidisciplinary team identified key change concepts. These were tested in a series of PDSA cycles. Interventions included: monthly diabetes dashboard review, care navigator (CN) outreach, physician engagement to allow timely scheduling, rescue visits slots added to nurse practitioner (NP) schedule, CN appointment reminder call, EHR health maintenance implementation, RN champion outreach telephone visit to review blood sugars and provide insulin adjustment.
Results: As of May 2022, the percentage of children with T1D who had four or more clinic visits per year increased by 35% from the May 2021 baseline of 53% to 88% (goal: 75%); Figure 1.
Conclusions: Continued follow‐up of this population will help identify barriers to clinic attendance and the interventions with the most impact. New strategies to address inequities and improve visit frequency in T1D are needed.
Keywords: pediatric, type 1 diabetes, visit attendance
Figure 1. Percent of Medicaid‐insured children with T1D with at least 4 clinic visits per year
Abstract #26 Improving continuous glucose monitoring use in adolescents and young adults
Faisal S. Malik; Samantha G. Perez; Sarah Lowry; Kathryn W. Weaver; Alissa J. Roberts
Department of Pediatrics, University of Washington, Seattle, WA; Seattle Children's Research Institute, Seattle, Washington, USA
faisal.malik@seattlechildrens.org
Background/Objective: A significant percentage of adolescents and young adults (AYA) with type 1 diabetes (T1D) do not use continuous glucose monitors (CGM) to support diabetes management. We sought to examine the impact of the University of Washington (UW) AYA Diabetes Program on CGM use among AYA with T1D and assess whether this varied by health insurance and race/ethnicity.
Methods: The UW AYA Diabetes Program is a partnership between Seattle Children's Hospital and UW Medicine designed to meet AYA healthcare transition and mental health needs. This study included patients seen between 2017–2021. We calculated proportions and 95% CIs using generalized linear models with a log link and robust variance estimator to cluster on individuals to account for repeated measures. We assessed for possible effect modification by health insurance and race/ethnicity.
Results: In this cohort of 526 patients (72% non‐Hispanic White, 27% public insurance), adjusted CGM use at their baseline AYA Diabetes Program visit was 64% (95% CI: 60%–69%). This increased to 88% approximately 24 months after program enrollment (95% CI: 82%–94%). There was evidence of effect modification by health insurance type for CGM use (p < 0.01) but not by race/ethnicity. Baseline‐adjusted CGM use was significantly lower among public vs private insurance participants (50%, 95% CI: 41%–59%; vs. 69%, 95% CI: 64%–74%) but not after 6‐months of program participation (77%, 95% CI:65%–89% vs. 80%, 95% CI: 74%–86%).
Conclusions: The UW AYA Diabetes Program proved to be successful in increasing diabetes technology and mitigating baseline health insurance disparities in AYA with T1D.
Keywords: adolescent, glucose self‐monitoring, health inequities, type 1 diabetes, young adult
Figure 1.
Abstract #27 Improving mental health screening in an adult endocrinology clinic
Deene Mohandas; Jacob Less; Marina Basina
Stanford University School of Medicine, Department of Medicine, Division of Endocrinology, Gerontology, and Metabolism Stanford, California, USA
deenem@stanford.edu
Background/Objective: People with type 1 diabetes (T1D) are at higher risk for depression and anxiety. Reports by JDRF indicate that one in four people with diabetes is impacted by depression. Routine screening is essential to reduce psychological comorbidities and improve diabetes care. This quality improvement initiative was implemented to increase screening for depression and anxiety from 0% to 50% over 12 months.
Methods: Through seven Plan‐Do‐Study‐Act (PDSA) cycles, adult T1D patients were screened using PHQ‐2 to 8 and GAD7. Surveys were assigned electronically one to 2 weeks prior to visits. Patients already followed by a mental health provider or screened in the past 6 months were not re‐screened. One to 2 days before their visit, patients were reminded to complete assigned questionnaires. Patients with elevated scores were provided with resources via secure messages and encouraged to discuss them during their visit.
Results: Following a series of rapid PDSA cycles, the screening rate increased from 0% to 56% over 6 months (Figure 1). During this time, the single endocrine provider had 256 unique T1D patient visits, where 211 patients were assigned surveys.
Conclusions: Self‐management is crucial in the treatment of T1D and is significantly impacted by an individual's mental health. Our initiative demonstrates the possibility of successfully introducing screening in a sensitive patient population. It highlights that additional studies are needed to standardize mental health and depression screening tools in adult endocrinology clinics.
Keywords: anxiety, depression screening, mental health, type 1 diabetes mellitus
Figure 1.
Abstract #28 Certified diabetes care and education specialists' perspectives on the 4 T program
Jeannine Leverenz, Piper Sagan, Anjoli Martinez‐Singh, Barry Conrad, Julie Senaldi, Annette Chmielewski, Brianna Leverenz, Priya Prahalad, David Maahs
Stanford Children's Stanford, California
jleverenz@stanfordchildrens.org
Background/Objective: The 4 T Program (Teamwork, Targets, Technology, and Tight control) at Stanford Children's aimed to intensify new‐onset type 1 diabetes (T1D) education. In the Pilot 4 T Study, HbA1c improved by 0.5% at 12 months compared to historical controls. Certified diabetes care and education specialists (CDCES) were key to the development and implementation of the program.
Methods: Youth with new‐onset T1D start continuous glucose monitoring (CGM) the first month of T1D diagnosis, and CDCES review CGM data weekly. The CDCES sends families electronic health record‐based education and dose adjustment messages. All CDCES who participated in the 4 T program (n = 10) received an email with a link to an anonymous six‐question RedCap survey to assess their experiences with the study.
Results: The survey was completed by 90% of CDCES. A third of CDCES (n = 3) felt that the process of starting CGM on all patients with new onset T1D and a 1‐week follow‐up added a burden to the workload of CDCES. Another 44% (n = 4) felt that weekly CGM reviews added to the workload of the CDCES. However, all CDCES felt that the 4 T Program supported the needs of patients and families, empowered CDCES, and was rewarding for the CDCES team (Figure 1).
Conclusion: While the 4 T Program added to the CDCES workload, they all felt that the program was beneficial to patients and families and rewarding for the CDCES team. Incorporating CDCES perspectives into program development can lead to successful program development and increased job satisfaction.
Keywords: continuous glucose monitors, pediatrics program development, type 1 diabetes
Figure 1. CDCES responses to the survey questions.
Abstract #29 A program to decrease diabetic ketoacidosis (DKA) admissions: diabetes wellness program (DWP)
Roberto Izquierdo; David W. Hansen; Margaret Greenfield; Emilie Hess; Christopher P Morley; Karen Kemmis; Beth Wells; Janine Robbins; Ann Marie Sanders; Hollie Cartini
SUNY Upstate Medical University Syracuse, New York, USA
izquierr@upstate.edu
Background/Objectives: Pediatric DKA admissions rose by 40% in the US from 2006 to 2016, with vulnerable subgroups having the highest risks. Non‐Hispanic Black, Hispanic groups, and those with public insurance have the highest rates of DKA. DKA admissions are costly (>$20 000 per hospitalization). The purpose of this quality initiative project was to identify youth at high risk for DKA, and the aim was to develop a diabetes program (DWP) with a set curriculum to prevent DKA.
Methods: We identified youth at risk for DKA admissions if they had two or more DKA admissions, frequent ER visits, sustained A1c > 14%, or frequent outpatient calls for hyperglycemia with ketonuria. We identified two intervention cohorts of 16 patients. Each group met with a Nurse Practitioner, Registered Nurse (RN), Registered Dietitian (RD), or Diabetes and Education Specialist (DCES) every 4–6 weeks, and a Social Worker as needed. The RN and RD DCESs followed a diabetes education curriculum created for this program. We measured A1c and DKA admissions pre‐ and post‐intervention. Participants were administered quality‐of‐life (QoL) and satisfaction questionnaires.
Results: DWP participants had significant decreases in A1c and DKA admissions compared to a control group (Not Enrolled) (Table I). Seventy‐seven percent of the participants who completed the program had improvement in QoL scores, and all were highly satisfied.
Conclusions: This DKA prevention program was effective in improving A1c and preventing DKA admissions in a group of youth with type 1 diabetes at high risk for DKA.
Keywords: diabetic ketoacidosis, quality improvement, type 1 diabetes
TABLE 1: Change in A1c and DKA admissions
| Completed DWP | Not Enrolled | p‐value | |
|---|---|---|---|
| A1c change (mean ± SD) | −1.49 ± 2.4 | −0.54 ± 1.98 | 0.05 |
| DKA admissions change (mean ± SD) | −2.39 ± 3.5 | −0.94 ± 1.62 | 0.009 |
Abstract #30 Nuances of socioeconomic status and health insurance associations with HbA1c in adults with type 1 diabetes
Caitlin S. Kelly; Huyen Nguyen; Katherine S. Chapman; Megan E. Peter; Kelsie LaFerriere; Julia Ravelson; & Wendy A. Wolf
T1D Exchange Boston, Massachusetts, United States of America
ckelly@t1dexchange.org
Background/Objective: Markers of lower socioeconomic status (SES)—such as personal income and education—are associated with higher HbA1c in people with type 1 diabetes (T1D), but they do not occur in isolation. We explored concurrent associations between demographics, SES markers, residential characteristics, and health insurance with HbA1c and potential moderating relationships of insurance type with county‐level income.
Methods: Participants were adults in the T1D Exchange Registry—an online longitudinal study of people with T1D—who completed the Baseline Registry Questionnaire. Health insurance was categorized into private (79.5%), Medicaid (9.3%), Medicare (9.4%), and no insurance (1.9%). We used hierarchical linear regressions to examine associations between health insurance type and residential characteristics (non‐Metropolitan residence, county‐level income) with self‐reported HbA1c after adjusting for diabetes technology use, demographics, and SES markers. We then examined whether having Medicaid (vs. private insurance) changed these associations.
Results: Participants (N = 9027) lived in metropolitan areas (85.5%) and average county‐level annual income was $77,710. Mean HbA1c was 7.3% (SD = 1.63). After adjusting for diabetes technology, demographics, and SES, participants with Medicaid (Β = 0.24) and participants in lower‐income residential areas (Β = −0.06) reported higher HbA1c (p < 0.001). Further, having Medicaid moderated the association between county‐level income and HbA1c. For those with Medicaid, as county‐level income increased HbA1c decreased (simple slope = −0.29, t = −4.30, p < 0.001).
Conclusions: Our results suggest a multifaceted approach to SES which includes community resources (i.e., median county income) may help to explain inequities in HbA1c.
Keywords: health insurance, socioeconomic status, type 1 diabetes mellitus
Abstract #31 Standardizing insulin pump therapy initiation in children with type 1 diabetes
Mili Vakharia1, Sarah Lyons1, Don Buckingham1, Daniel Desalvo1, Rona Sonabend & Grace Kim
Division of Pediatric Diabetes and Endocrinology, Department of Pediatrics, Baylor College of Medicine/Texas Children's Hospital, Houston, Texas, USA
vakharia@bcm.edu
Background: Advancements in diabetes technologies, such as insulin pump therapies, can lead to improved glycemic control, reduced hypoglycemia, and enhanced quality of life. The International Society for Pediatric and Adolescent Diabetes and American Diabetes Association supports the need for structured technology education for patients with type 1 diabetes mellitus (T1D) to optimize success with device use. However, the underutilization of insulin pumps and general prescribing barriers have been noted at our institution. We implemented a QI initiative aimed at standardizing and increasing insulin pump use in all new‐onset T1D patients less than 1 year from diagnosis by 5%.
Methods: A series of Plan‐Do‐Study‐Act cycles were implemented beginning in January 2021 at one community campus, including:
Educating the diabetes educators and providers about the new onset pump initiation protocol, safe start criteria, pump action plan, and pre‐ and post‐pump visit checklist.
Implementation of new processes, including early introduction of pump technology and education to patients/families at 90 days from initial diagnosis.
Scheduled post‐pump follow‐up visit with provider and diabetes educator.
Results: In patients with a T1D duration of <1 year, insulin pump usage has increased from a baseline of 20% in January 2021 to 30% in June 2022 and remains sustained (Figure 1). Given this improvement, we are spreading the initiative to additional campuses.
Conclusion: Standardizing the approach to initiating insulin pump therapy, including providing a multidisciplinary pump education, facilitates a structured way to increase the uptake of pump technology in patients with T1D. It may help address healthcare disparities through the elimination of unconscious provider prescriber biases.
Figure 1.
Abstract #32 “TechQuity” and peer support to reduce disparities in glycemic outcomes in children with type 1 diabetes (T1D)
Jenise C. Wong; Barbara Liepman; Katie Craft; Diana Arellano; Mackenzie Allen; Kathy Love; Annapurna Vishnubhotla; Angel Nip
Division of Endocrinology, Department of Pediatrics, University of California San Francisco San Francisco, California, USA
jenise.wong@ucsf.edu
Background/Objective: Disparities in glycemic outcomes exist between publicly and privately insured children with T1D. During fiscal year (FY) 21, publicly insured patients with T1D at Benioff Children's Hospitals (East and West Bay) had a 1.3% higher median hemoglobin A1c (A1c) than privately insured patients. The goal was to reduce the monthly median A1c gap to ≤1.24% (5% reduction) between publicly and privately insured children with T1D by the end of December 2021, and further reduce it to ≤1.17% (10% reduction) in April–June 2022.
Methods: A multidisciplinary Task Force was established in May 2021 to identify, test, and implement change ideas. The Task Force hypothesized that “TechQuity” (the strategic development and deployment of technology to advance health equity) and peer support were key drivers to achieving A1c goals.
Results: An educational handout on continuous glucose monitoring (CGM) was created and made available to staff to distribute to patients not using CGM. The patient voice was captured by the survey, and patients identified peer support to help with diabetes management. The Task Force worked with Diabetes Youth Families, a local nonprofit, to establish a peer support “buddy” program. By the end of FY22, the cumulative median A1c gap decreased to 0.9% and 1.1% in the East and West Bay clinics, respectively (Figure 1A and B).
Conclusions: Establishing a Task Force, increasing CGM education, and creating opportunities for patients to connect with peer support contributed to a reduction in disparities in A1c. Future interventions will support increased CGM use and outreach to patients with A1c levels above the goal.
Keywords: child; diabetes mellitus, health equity, quality improvement, type 1; glycated hemoglobin
Figure 1. Run Chart of Gap in Median A1c between Publicly and Privately Insured Children with T1D in (A) East Bay Clinics and (B) West Bay Clinics
Abstract #33 Improvement in screening for depression in adolescents with diabetes
Barbara Liepman; Angel Nip; Jenise C. Wong
Department of Quality and Patient Safety and Department of Pediatrics, Benioff Children's Hospital, University of California San Francisco San Francisco, California, USA
barbara.liepman@ucsf.edu
Background/Objective: Depression is common among adolescents, but rates increase significantly with diabetes, and is associated with suboptimal glycemic outcomes and increased diabetes‐related complications. Annual depression screening of adolescents is recommended as a part of standard care. The screening rate of adolescents seen in the pediatric diabetes clinics at Benioff Children's Hospital (BCH) was 29% in 2019. The aim of this project was to provide standardized psychosocial screening and achieve screening rates of ≥40% in fiscal year (FY) 21 and ≥ 50% for at least 9 of 12 months in FY22 for adolescents 13–17 years with diabetes at all diabetes encounters at BCH.
Methods: A multidisciplinary Task Force was established in February 2021. Clinic workflows were created to screen adolescents with diabetes using the validated PHQ‐9 screening tool at least annually at in‐person visit encounters at the two main diabetes clinics at BCH (East and West Bay). Patients with positive screens were referred to a social worker for further mental health assessment.
Results: The depression screening workflow was initiated in March 2021. In FY22, the cumulative depression screening rate increased to 56%, which was a 27% increase from the 2019 baseline (Figure 1). Adolescents with specific mental health needs and challenges were identified and referred for appropriate evaluation.
Conclusions: Systematic depression screening at in‐person clinic visits can be reliably implemented. As a result of the screenings, adolescents with diabetes needing mental health evaluation were identified and referred to appropriate resources. Future considerations include expanding screening to telehealth visits and satellite clinics.
Keywords: adolescents; diabetes mellitus; type 1; type 2; depression; psychosocial screening
Figure 1. Run Chart of Depression Screening in Adolescents with Diabetes
Abstract #34 Pediatric type 1 diabetes technology use: hispanic caregiver's perspective
Makacio Morillo, Mariaester, Hernandez Lisandra, Sanchez Janine
Jackson Memorial Hospital, University of Miami Miami, Florida, USA
mariaester.makaciom@jhsmiami.org
Background/Objective: Studies have shown that Hispanics have low use of insulin pumps and continuous glucose monitors (CGM) when compared to Whites. (Agarwal, S, 2021, Apr). Racial disparities in technology use and diabetes outcomes persist regardless of insurance status. (Lipman, TH, 2021, Mar). Our aim was to obtain a better understanding of barriers that keep people from using advanced technology and prevent health disparities.
Methods: We completed semi‐structured interviews and two focus groups with a sample group of Hispanic parents of type 1 diabetes (T1D) patients. We focused our interview on four main topics: sensor usage, pump vs. smart pen vs. MDI, training via video conferencing versus in‐person, and suggestions to improve care. We compared the results to study what we did with Black parents.
Results: We gathered data from 9 Hispanic mothers of T1D patients. The interview format was chosen by parents—22% individual interview, 11% email, and 66% focus group. All parents preferred interviews in Spanish. 100% reported using sensors regularly, 33% insulin pump, 55% smart pen, and one patient used neither. Key barriers to technology use were fear of doing something wrong at the beginning, difficulty with initial telemedicine training during COVID, and insurance problems, including language barriers and refills.
Conclusions: Overall, when compared to our Black focus group, we received similar results. They are satisfied with the technology and feel it increases their understanding of diabetes. Insurance plays a limiting role. More studies are needed to increase understanding of our population and increase device use.
Keywords: health disparities, technology use, type 1 diabetes
Abstract #35 Overcoming barriers to increase CGM use for youth with type 1 diabetes
Elizabeth A. Mann, Whitney N. Beaton, Rachel J Fenske, M. Tracy Bekx
University of Wisconsin School of Medicine and Public Health & UW Health Kids Madison, Wisconsin, USA
eprange@wisc.edu
Background/Objective: The use of continuous glucose monitors (CGM) has been shown to reduce hemoglobin A1C and improve diabetes outcomes in youth with type 1 diabetes (T1D). We led a quality improvement initiative to increase the percentage of all youth with T1D in our practice who initiated the use of CGM from 64% to 75% between January 2021 to May 2022 and to increase the percentage of youth with T1D using a CGM at 3‐months post‐diagnosis from 68% to 80% between September 2021 to May 2022.
Methods: We identified key drivers to increase CGM utilization, which included early initiation of CGM, streamlining the ordering process, and addressing CGM barriers for public insurance. Using PDSA cycles, we incorporated a CGM education protocol into new onset education, clarified the ordering process internally and through partnerships with local CGM and DME representatives, and advocated for reducing barriers to CGM access with our state Medicaid program. Monthly reports of measures were generated and reviewed using run charts or SPC charts.
Results: P SPC chart showed special cause was achieved in May and November 2021, increasing the percentage of youth with T1D using CGM from 63.6% to 67.3% to 73.1% (Figure 1A). P SPC chart showed no special cause was achieved in percent using CGM 3‐months post‐diagnosis, though the last 5 data points above the median suggest a potential future shift.
Conclusions: Rates of CGM use in youth with T1D increased through improvement efforts focused on standardized education and ordering and addressing public insurance barriers.
Keywords: adolescent, child, glucose sensor, hemoglobin A1c, type 1 diabetes mellitus
Figure 1. A. P SPC chart showing percent of youth diagnosed with type 1 diabetes (T1D) for more than one year who are using a continuous glucose monitor (CGM). Special cause achieved in May 2021 and November 2021. B. P SPC chart showing percent of youth newly diagnosed with T1D using a CGM three months after diagnosis.
Abstract #36 Scholarly dissemination from the T1DX‐QI network, 2020–2022
Nicole Rioles; Holly Hardison; Shivani Agarwal; Shideh Majidi; Osagie Ebekozien
T1D Exchange, QI, and Population Health, Boston, Massachusetts, USA
nrioles@t1dexchange.org
Background/Objective: The T1D Exchange Quality Improvement Collaborative (T1DX‐QI) has expanded its publications scope and impact through a clear and intentional process to increase authorship opportunities and dissemination. T1DX‐QI works with interdisciplinary diabetes care teams to build quality improvement capacity and perform population health analysis.
Methods: As part of the publication process, T1DX‐QI made calls for abstracts and organized manuscript writing teams to increase authorship opportunities for junior faculty and disseminate learning on cross‐disciplinary topics to improve diabetes care. First, authors were chosen based on their subject interest, knowledge, and capacity. We analyzed T1DX‐QI publications from January 1, 2020 to June 30, 2022, tracking conference type, journal type, journal impact factor, and the number of citations for articles published in the study period.
Results: From 2020 to 2022, T1DX‐QI published 32 articles across 12 journals, with impact factor ranges of 1 to 19. Eight articles were published in 2020, 16 in 2021, and eight as of June 2022. T1DX‐QI presented 13 abstracts in 2020, 41 in 2021, and 21 in 2022 across six international conferences. Of the 32 publications, 15 (47%) were first authored by new investigators, and 19 (60%) were from pediatric centers.
Conclusions: T1DX‐QI has purposely created publication opportunities, supporting first‐time authors and junior faculty to co‐author with senior faculty for mentorship and professional development. Furthermore, by publishing T1DX‐QI's work, more practicing clinicians can now learn improvement insights that continue to build the dissemination potential of successful clinic‐level interventions to improve diabetes care for people with T1D.
Keywords: diabetes, learning health system, publishing, quality improvement
Fig 1. Articles and Abstracts Published