Diabetes On Track: Redesigning Diabetes Management in Rural Primary Care Free

After a thorough community mapping process, two communities were selected to be project sites. For a community to be eligible, at least one primary care clinic within the community had to be using the same electronic health record (EHR) system as Nebraska Medicine, and the community had to meet the following predetermined rural definitions: one had to be a small rural community, defined as having a population of 2,500–5,000 and being nonadjacent to a metro area, with low commuting; the other had to be a large rural community, defined as having a population of 20,000–25,000 and being adjacent to a metro area, with high commuting.

All health care practices within each identified community were approached and invited to participate. Conversations occurred with administrators at the health system level and at the clinic level. In community A (population 24,691, with 11.7% living in poverty), out of four primary care clinics, two (both hospital-affiliated) chose to fully participate, one agreed to share their electronic clinical quality measure data, and the other declined to participate. At the intervention's start, the two participating clinics merged and became one larger clinic (hereafter referred to as clinic A). In community B (population 5,820, with 19% living in poverty), the lone primary care clinic in the community agreed to fully participate (hereafter referred to as clinic B).

Each participating clinic’s practice manager was the point person for the Diabetes On Track initiative. They worked closely with clinic nurses, physicians, and administrative staff to implement solutions. There was also a physician champion identified for each clinic.

A multidisciplinary project team consisting of two endocrinologists, a general internal medicine physician, a statistician, a registered dietitian/certified diabetes care and education specialist (RD/CDCES), a registered nurse (RN) clinical project coordinator, and a nurse scientist was based at the University of Nebraska Medical Center (UNMC). This team supported the primary care clinics as they tried to institute practice changes to improve diabetes-related outcomes. In addition, clinical and project team members pulled in additional resources as needed, such as information technology (IT) consultants to make functional changes to the EHR system and clinical pharmacists to help establish new protocols.

Once the communities were identified and data use agreements were signed, the project team met with the clinic sites and asked the clinics to complete an informal assessment of its strengths and weaknesses (Supplementary Material). Focusing on assets, clinic staff worked with the project team to restructure existing resources and use the initial funding to establish practices that would be self-sustaining at the end of the funding period.

The initial intervention bundle was co-developed by the Diabetes On Track research team and each clinic team based on a literature review, suitability for the specific site, and expert opinion. Of all the possible interventions, only the Project ECHO (Extension for Community Healthcare Outcomes) sessions were required as a condition of being involved in the project. The rationale was to provide both sites with the same education about standards of care and to encourage them to interact with each other and share rural-specific experiences. Each clinic’s staff and administrative team controlled which strategies they would test at their site and could tailor the strategies as they found appropriate. In addition, they had complete freedom to initiate other strategies they thought would be helpful. The strategies and interventions offered to each clinic site are discussed in detail below.

Project ECHO is an evidence-based framework that has been used for the past 20 years as a virtual resource for clinicians to engage with their peers in an “all teach, all learn” environment (19). Project ECHO sessions feature a short didactic component followed by a de-identified case presented to a multidisciplinary expert team and other participants for discussion about recommended courses of action. These sessions are led by the UNMC RN clinical project coordinator.

The Association of Diabetes Care & Education Specialists (ADCES) is a leading association for diabetes education. Its membership benefits include continuing education, technology and reimbursement resources, patient education resources, and access to case studies.

The UNMC College of Nursing offers a year-long fellowship for RNs who want to advance their primary care skills. This distance-learning model includes self-guided online modules as well as monthly meetings, totaling >30 hours of continuing education. The fellowship is intended to prepare RNs to successfully pass the Ambulatory Care Nursing Certification exam of the American Nurses Credentialling Center.

Interested clinic staff were invited to participate in becoming a certified health coach, earning the CDCES certification, and/or earning the Board Certified in Advanced Diabetes Management designation. All of these additional credentials have billing potential, thus providing a mechanism for sustainability. The availability of a health coach in the primary care setting has been shown to improve A1C, systolic blood pressure, and physical activity levels (20).

Knowing that multidisciplinary care is a critical component of success (3), funding was made available for part-time pharmacy support to collaboratively manage medication titration, review potential medication interactions, and assist with laboratory monitoring (21,22).

A standardized, widely used hypoglycemia protocol was offered to provide a consistent process for the immediate treatment of emergent hypoglycemia episodes in adult patients in the ambulatory clinic setting.

Re-investment of shared savings revenue is an ongoing source of funding for population health initiatives such as improved diabetes team-based care. Consultations with university health system population health leaders were offered to each clinic to assess its readiness to change, stage in the value-based journey, and potential for long-term success and sustainability.

This EHR system function allowed clinicians to see how their patient panel was performing on the following diabetes-related quality metrics: A1C measurement, A1C control, nephropathy screening, retinopathy screening, blood pressure control, lipid monitoring, and statin therapy. Physicians could view their panel in real time to see if they were meeting, exceeding, or falling short of established targets for each indicator. These dashboard targets had been previously established by the EHR system. Clinics were later offered incentives semi-annually for meeting their targets or stretch goals for each metric over a 2-year period. The incentive program used target and stretch goals based on national benchmarks and/or top quartile performance regionally, with efforts aimed at improvement over time.

Because the project was iterative in nature, additional strategies brought forth by the practices were developed and implemented.

After clinic assessments were conducted and initial interventions were chosen, the RN clinical project coordinator and the RD/CDCES held a weekly videoconference meeting with clinic staff to discuss implementation of the intervention components and troubleshoot any issues that may have arisen. This iterative exploratory process allowed clinic staff to adjust any interventions to best support the goal of improving diabetes outcomes.

In addition to these video meetings, robust technical assistance was made available to help with implementing the specific interventions. For example, when a clinic chose to integrate a provider dashboard, UNMC personnel experienced in dashboard setup and functionality worked closely with the clinic to ensure that the intervention would be used as proposed.

Site visits occurred twice per year to provide face-to-face technical assistance, particularly with the providers who may only be tangentially engaged in the project's daily implementation. This opportunity allowed for open discussion about project progress, sustainability, and future work among stakeholders.

The research team defined six core measures to track progress over time (Table 1). These measures were derived from annual aggregate medical record data from the participating clinic in each community. The 18-month time frame for this analysis was from July 2022 to January 2024.

Once the two clinics chose strategies, the team developed process metrics to assess progress toward the provision of robust clinician education and enhancement of patient self-management by establishing a team-based care model.

Baseline and 18-month findings for the six core metrics are shown in Table 2. The prevalence of patients at risk for prediabetes based on age and BMI was similar for both clinics, but screening rates of these at-risk individuals was higher in clinic A. There was an increased trend in screening rates for prediabetes at 18 months in both communities. Clinic A had a higher percentage of individuals with confirmed prediabetes or diabetes than clinic B. The percentage of people with diabetes with an A1C >9% declined in both clinics over 18 months. Among those with an established diagnosis of type 2 diabetes at both clinics, there was an increase in documentation of A1C in the EHR system. While we will continue to track these outcome measures for the next 18 months, it is also important to highlight progress on the different activities the clinics are engaged in to understand whether and how these interventions may be contributing to changes in outcome metrics.

In addition to the required participation in Project ECHO sessions, clinic A opted to engage an RN health coach, initiate pharmacy co-management, expand diabetes education training opportunities, and use the provider dashboard with incentives. Between June 2022 and January 2024, there were 23 ECHO sessions, with an average of 10 individuals from clinic A attending. Over 18 months, the RN health coach had 163 referrals and met with 135 unique patients, with 19% of those patients having more than one visit. An ambulatory referral order to a health coach was added to the EHR system in spring 2023 to help track and facilitate the referral process within the clinic. The inpatient pharmacy team assigned a pharmacist to spend 1 day per week in the clinic setting to co-manage patients with type 2 diabetes, with 19 referrals placed by primary care providers over 18 months. To further expand the multidisciplinary team for diabetes care and management, a diabetes educator within the local hospital system opted to work toward CDCES certification, and the RN health coach completed the ADCES7 Self-Care Behaviors training. Trends in the provider dashboard are summarized in Table 3. The largest improvements were seen in nephropathy and retinopathy screenings. Clinic A elected to directly incentivize its clinicians; 50% of the incentive revenue was divided among clinicians, and the remaining 50% was placed in a general fund for clinic use.

Participating in the initial Diabetes On Track interventions led to the organic development of other strategies to improve care. One change was the development of a call list of patients at risk for prediabetes (based on age and BMI) who had not been screened in the past 3 years. Originally conceived as a pre-visit planning call, the intervention expanded to reaching patients at highest risk who had not been screened and asking them to come to the clinic for laboratory testing regardless of whether they had an upcoming visit scheduled. Since its implementation in March 2023, 198 additional patients have been contacted, and the screening rate for people at risk for prediabetes has increased from 38 to 49%. To increase diabetic retinopathy screenings, clinic A purchased a retinal camera that detects diabetic retinopathy through the use of artificial intelligence. Additionally, the hospital system affiliated with clinic A made the decision to join an accountable care organization in the summer of 2023.

In addition to participating in the Project ECHO sessions, clinic B chose to create an RN health coach position (one fulltime position shared by two RNs), take advantage of the pharmacy mentorship, increase the number of CDCESs on staff, and use the provider dashboard with incentives. Average attendance from clinic B across the 23 Project ECHO sessions was 3.3 individuals per session. This most commonly included clinic B’s physician champion and the two RNs sharing the full-time health coach role. The RN health coaches began to see patients in January 2023 and established an EHR system ambulatory referral to health coaching in the fall of 2023. Over 15 months, they received 74 referrals, resulting in 67 unique patient visits, with 24% of those patients having more than one visit. In addition, the RN health coaches received pharmacy support to set up continuous glucose monitoring (CGM) for eligible patients. Before the project, CGM training was not offered by clinic B, but by January 2024, six patients had been trained on the use of personal CGM. Both RN health coaches completed the ADCES7 Self-Care Behaviors training and are pursuing their CDCES certification. In addition, they encouraged local community health workers (CHWs) to complete the CHW track of the ADCES7 Self-Care Behaviors training. Trends in the EHR system provider dashboard metrics are shown in Table 3. Similar to clinic A, the largest changes were seen in nephropathy and retinopathy screening rates. Of note, clinic B chose not to directly incentivize their clinicians for achieving target metrics, and all incentive revenue went to a general clinic fund.

Baseline screening rates were low in both communities, particularly in clinic B. These low screening rates suggest that there are likely more undiagnosed individuals in the community with either prediabetes or type 2 diabetes. The rates of prediabetes in both communities are similar to the national average of 33% (23). The national rate of adults diagnosed with type 2 diabetes is estimated to be at 11.3% (23). Although clinic B’s rate is lower than this national average, it may be the result of its low screening rates.

There are several potential reasons for the improvement seen in provider dashboards over time. First, a laboratory mapping issue was resolved, which was particularly noticeable in the nephropathy screening rates in clinic A. Additionally, several providers used the real-time feedback of the dashboard as motivation to improve screening rates, asking nursing staff to follow up with patients who had not yet completed their recommended screenings. Finally, in clinic A, individual providers were directly incentivized to meet dashboard targets, which we believe was a likely contributor to their greater success relative to clinic B. As clinics have seen the value of the dashboard in meeting quality metrics, both clinic mangers have voiced the intention of sustaining this intervention (i.e., contracting with UNMC to continue providing the service) after the project period ends.

Recognizing the rural mortality penalty, rural health systems are aware of the need to improve diabetes outcomes in their patient populations. Many of the efforts to date are directed at implementing a specific evidence-based program or care path. The provision of team-based care is one strategy that has been trialed in many rural clinics (22,24–26). For example, Hurst et al. (22) describe a team-based approach using pharmacy and health coaching professionals to foster the provision of diabetes management care and education at a free rural Midwestern clinic, resulting in improved LDL cholesterol, blood pressure, and A1C levels. Although implementing team-based care can clearly be effective, it could be challenging in resource-constrained settings to enlist and coordinate such a team.

Virtual care is another strategy often implemented in rural clinics, particularly to connect with diabetes specialists. Telehealth is an effective resource in many situations, but researchers note that the same disparate groups that have trouble accessing care also have trouble accessing virtual care (27,28). As Eiland and Drincic (29) note, virtual care has to be implemented thoughtfully, taking into consideration the community context and its associated constraints.

Instead of focusing on the implementation of a particular intervention, Diabetes On Track was designed to provide rural primary care clinics an opportunity to develop, test, and adapt their own interventions. Focusing on the process of changing systems has three distinct advantages. First, local decision-makers become more aware of the strengths and limitations of their own practice environment, particularly after conducting a rapid needs assessment. Therefore, the local practices have a better idea about which interventions are likely to be successful at improving patient care and increasing opportunities for billable services. Second, the process is designed to give clinics access to real-time local data to inform their decision-making. For example, noting the percentage of at-risk patients who had not been screened for prediabetes or type 2 diabetes, clinic A decided to create an EHR system worklist and call those individuals to come to clinic for screening. Third, because this is a locally driven effort, clinics have the freedom to innovate. Instead of working harder doing things the same way, the practice managers, clinicians, and staff can decide to shift strategies. For example, to increase referrals to the RN health coach, the IT teams added an ambulatory patient referral order to the EHR system in both clinics.

Studies have indicated how important it is for clinics to have autonomy when deciding on the best approaches to care and to have the freedom to innovate (26), but this goal has not been realized often in practice (30). Eighteen months into this project, there seem to be differences in activities and outcomes between the two clinics. It is possible that the menu of interventions was not sensitive enough to the unique constraints of the smaller clinic B. Most providers at clinic B also provide care at one or two other satellite clinics in the area. Given this fact, there may be a limited sense of ownership regarding how this clinic functions. Kavanagh et al. (31) posit that “soft infrastructure” (e.g., culture, trust, and identity) is just as important as material resources. Clinic B may need to focus on its soft infrastructure before it can develop or implement interventions that produce more robust results.

Several challenges have arisen in implementing this initiative. Health care has become so acculturated to top-down decision-making that there are both real and perceived barriers for clinic staff to feel that they have decision-making power with this program. Sometimes, the locus of decision-making does reside further up the vertical chain of command to the health system that oversees the clinic (e.g., the decision to directly incentivize providers). In these cases, the project team based at UNMC has helped to facilitate these processes.

Another limitation is that the Diabetes On Track health care system team has not focused on the patients’ voice and changes patients would like to see in their health care system. The community system team has included patient perspectives of the health care system in their community coalition efforts, but future work will be needed to translate this perspective into actionable interventions within the health care system.

Finally, this case study design cannot specifically attribute any changes in outcomes to one intervention. However, this initial phase of Diabetes On Track is considered a feasibility trial to determine whether it is possible to engage clinic staff in developing local solutions to improve diabetes care. A future community-randomized trial will be needed to rigorously evaluate whether locally driven system changes result in improved outcomes.

Historical approaches to improving diabetes outcomes in rural primary care settings have generally been developed with a top-down structure. Because of the unique constraints found within rural primary care settings (e.g., unstable staffing, financial constraints, and fewer specialty care resources), a more effective approach to improving outcomes may include giving primary care clinics the autonomy to use local data to develop and implement their own innovative strategies.

This work was funded, in part, by the Diabetes Care Foundation of Nebraska.

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

K.L.P. contributed to the investigation and wrote the manuscript. L.A.E. and S.M.M. contributed to the investigation and methodology. S.N. and S.S.K. contributed to the investigation. C.S.W. handled data curation and formal analysis. C.D. was responsible for project supervision and administration. All authors reviewed and edited the manuscript. C.D. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.