Mississippi is consistently ranked among the states with the highest rates of diabetes (n=326,420, 14.4%) and prediabetes (n=814,000, 35.6%) in the United States [24]. It is also the only state where every county is included in the Diabetes Belt, a cluster of counties in the Southeastern United States where obesity and physical inactivity account for one-third of all diagnosed diabetes cases [25]. Mississippi has the highest percentage of Black (n=1,111,340, 37.8%) populations of any state in the United States and persistently has higher rates of poverty (n=570,371, 19.4%) and a lower median household income (US $49,111) compared to national statistics (n=38,661,356, 11.6% and US $69,021, respectively) [26]. The exceedingly high rates of poverty contribute to Mississippi having the highest Federal Medical Assistance Percentage in the nation (77.27%) [27]. It is the nation’s fourth most rural state [20], where 65 (79.3%) of the 82 counties are considered rural areas. According to the Health Resources and Services Administration [28], every one of Mississippi’s 82 counties is designated a medically underserved area [28].

The University of Mississippi Medical Center (UMMC) in Jackson, Mississippi, is the state’s only academic medical center. Annually, UMMC physicians conduct over 535,000 clinical visits; 195,000 outpatient and 135,000 emergency room encounters; and 31,000 hospital admissions serving pediatric and adult patients statewide. The UMMC is a minority-serving and research-intensive health care organization with a state mandate to provide no less than 50% of its services to indigent persons (Miss. Code Ann. § 37-115-27; 2017).

The UMMC is also home to 1 of only 2 telehealth centers of excellence in the country funded by the Health Resources and Services Administration (6-U66RH31459-04-01). In partnership with telehealth, the UMMC launched its DPP program in 2020. Implementing the National DPP LCP through health care systems, such as the UMMC, has the potential to increase program reach and accessibility for residents who are among those at the highest risk for diabetes in the most medically underserved communities; however, no studies of the translation of the DPP have been conducted in Mississippi. The proposed study will be the first to comprehensively evaluate the translation of the National DPP and Medicare DPP in Mississippi.

The implementation plan was developed using the CFIR (Consolidated Framework for Implementation Research) [29]. The CFIR is a determinants framework for navigating the contextual factors that affect the uptake of innovations by health care institutions. The framework is comprised of 39 constructs across 5 complex interacting domains such as intervention characteristics, organizational inner setting context, outer setting context, characteristics of individuals involved with implementation, and implementation processes. The 4 interrelated processes include planning, engaging, executing and reflecting, and evaluating. The processes interact dynamically to achieve organizational (eg, UMMC health care system) and individual (eg, UMMC clinical providers) level use of an intervention as intended (eg, use of electronic health record system for screening and referral of patients at risk for diabetes). The UMMC’s a priori implementation plan includes 23 [24] discrete strategies described in Table 2 [30]. One strategy is taking a phased approach to implementation. Thus, in phase 1 of implementation, 3 clinical specialties were targeted based on provider buy-in including family medicine (n=8 clinical sites), lifestyle medicine (n=1 clinical site), and internal medicine (n=5 clinical sites).

The National DPP is a 12-month LCP delivered in a group setting by a trained lifestyle coach. The core 16 sessions are delivered weekly in the first 6 months of the program. In the maintenance phase, a minimum of 6 sessions are delivered monthly. The evidence-based program aims to develop behavior change skills leading to increased physical activity (150 minutes per week) and modest weight loss (7% reduction in total body weight), which are proven to prevent or delay diabetes in adults [7]. The DPP LCP is a clinical service of lifestyle medicine in the Department of Preventive Medicine in both the School of Medicine and John D. Bower School of Population Health at the UMMC. The LCP achieved full recognition from the CDC within the first year and, in 2020, became the third Medicare-approved DPP provider in Mississippi.

Lifestyle coaches are UMMC employees and include registered dietitians, project coordinators in population health, and community health workers. All interventionists completed a CDC-recognized lifestyle coach training, have a registered National Provider Identifier, and completed continuing education units. Newly trained lifestyle coaches shadow an experienced coach prior to leading an independent cohort of patients.

Group-based sessions are delivered using BigBlueButton, an open-source web-based classroom software [31]. This educational platform provides a unique link to connect participants using videoconferencing. Participants click a homeroom link, type their names, and enter the room. Weekly weight is tracked using digital body weight scales (developed by BodyTrace, Inc) equipped with cellular connectivity. Weekly physical activity minutes are reported to the lifestyle coach via phone call or email before or immediately following each session. Participant attendance, weight, physical activity, and billing (Medicare-eligible participants only) are tracked in patients’ electronic medical records in Epic (Epic Systems Corporation).

UMMC adult patients (≥18 years) who are overweight (BMI≥25 to 29.9 kg/m²) or obese (BMI≥30 kg/m²), have an HbA_1c between 5.7% and 6.4%, and are Medicare or Medicaid beneficiaries are recruited for the LCP by physician electronic referral. Using the Healthy Planet coordinated care management tool in Epic [32], physicians are prompted at the patient encounter by Best Practice Alert to diagnose, inform, and refer patients with prediabetes to the LCP. Epic is a patient-centered, cloud-based electronic health record software program comprised of multiple systems to facilitate patient care. Patients with 1 or more of the following are excluded: <18 years; pregnant or planning to become pregnant within the next 12 months; a primary or secondary diagnosis of diabetes, including all International Classification of Diseases, Tenth Revision (ICD-10) codes with a prefix of E11; or end-stage renal disease (ICD-10 code N18.6). Patients with a diabetes diagnosis and an HbA_1c value above 6.4% will be excluded.

Given that the LCP is delivered in a telehealth setting, it is preferred that patients have their own computer, tablet, or smartphone equipped with a camera and connected to a reliable Wi-Fi network. However, in a largely rural and underresourced state, this is not a realistic expectation. Thus, participants are eligible to participate if they own or have access to a computer, tablet, or smartphone connected to a reliable Wi-Fi network, have a mobile or landline phone to call into group sessions, or agree to use a study-provided tablet equipped with Wi-Fi and cellular connectivity. Study-provided tablets have restricted use and are provided to participants for the duration of the LCP and must be returned to the study team at the end of the program.

Electronic referrals to the LCP were extracted from Epic monthly between 2019 and 2021. In 2022, a dedicated LCP staff person was hired, and referrals have since been extracted from Epic twice weekly. Prior to making contact, the coordinator confirms the patient’s eligibility. Medical records of ineligible patients are updated, and they are referred to the Lifestyle Medicine clinic rather than being contacted. Among those eligible, up to 3 attempts are made to contact each patient including phone calls, voicemails, and emails. The date, time, type and outcome of contact, and the staff person’s name are documented. At a failed third attempt to make contact, the contact attempt details are recorded in the patient’s medical record and the referral is closed. Patients declining participation are also noted in their respective medical records and the referral is closed. Interested patients are added to a cohort and provided with information to attend session 0, an informational session.

Following session 0, interested participants are enrolled in the LCP and begin session 1 the following week. Scheduling cohorts consider patient preference and lifestyle coach availability. All session dates over the 12-month intervention period are determined prior to session 0, and participants receive the 12-month schedule at the start of the program. Makeup sessions are provided as needed but participants are encouraged to attend regularly scheduled group sessions. There is no limit to the number of participant makeup sessions. Up to 2 reminder contacts are made with each participant prior to each session. All participants with an email address are contacted via email, followed by a reminder phone call at which point participants are asked to report their weekly weight and physical activity minutes. For participants who do not have an email address, up to 2 phone call contacts are made. These procedures are consistent across participants and cohorts. In addition, when a participant misses 2 consecutive sessions, the project coordinator attempts to make contact to offer support in overcoming barriers and re-engaging them in the program. After 4 consecutive missed sessions and failed attempts to make contact, the participant is excluded from the program.

Implementation outcomes are mapped across the 5 domains of the RE-AIM framework (Table 2). The evaluation seeks to determine the intervention reach among patients at risk for diabetes, clinic and provider adoption, implementation fidelity, intervention effectiveness in preventing diabetes, and the organizational and patient uptake of the intervention. Implementation outcomes, research questions, measures, and methods are outlined in Table 3. Data sources will include electronic medical records, session notes provided by lifestyle coaches, intervention cohort and delivery details, meeting notes, and evaluation of DPP sessions conducted by an external reviewer.

The extent to which the target population is exposed to the intervention or population reach will be determined by the number of eligible patients referred to and enrolled in the LCP, yielding the appropriate referral to enrollment conversion rate. The a priori conversion rate goal is 50%. Reasons for exclusion and participant refusal to participate will be documented where possible. Descriptive characteristics such as age, sex, race, ethnicity, and BMI be explored and compared between patients eligible for the program who were referred and not referred.

The LCP effectiveness will be determined by the differential incidence of diabetes. Diabetes will be defined by diagnosis code or HbA_1c values greater than 6.4%. Data will be extracted from patient medical records and reviewed biannually. The a priori reduction in diabetes diagnosis among the intervention compared with the control group is 50%.

The adoption of the LCP will be defined as the number and proportion of clinical sites with referring providers making referrals to the LCP, and the number and proportion of eligible providers making referrals. Data will be extracted from medical records and reviewed quarterly. The a priori adoption clinical and provider addition rates are 70%.

Implementation fidelity and costs will be evaluated. Fidelity will be based on the UMMC’s recognition category by the CDC. The CDC standards consider participant recruitment, retention, attendance, the proportion of participants meeting the LCP 7% weight loss goal, the proportion of participants achieving 150 minutes or more of physical activity per week, and reductions in HbA_1c. Barriers, solutions, and facilitators will be documented by the project coordinator and lifestyle coaches in a weekly session log, during quarterly lifestyle coach meetings, and in strategic planning meetings with UMMC leadership and the American Medical Association Improving Health Outcomes team.

Ongoing organizational and participant engagement in the LCP will provide a robust measure of program maintenance. Longitudinal changes in the number of clinical sites and providers making referrals will be extracted from Epic quarterly, including data to monitor referral maintenance across sites and providers. Attendance and retention in the LCP will provide a measure of maintenance among participants. The a priori participant retention rate is 60%.

Study participants are identified in Epic using the eligibility criteria. Patients are placed in the intervention or comparison group based on natural selection. Each participant is assigned a study identification number that is matched to their medical record number. There is no interaction between study personnel and participants.

The intervention group includes eligible patients referred to and enrolled in the LCP. Once a participant is enrolled in the study, they are assigned a study identification number.

The comparison group includes patients meeting the inclusion criteria who are not enrolled in the LCP. Once a patient is enrolled in the DPP, they are no longer included in the control group. These participants include patients with and without a referral. Patients referred to the LCP, deemed ineligible based on insurance type, and referred to lifestyle medicine are excluded from the comparison group. Participants in the comparison group are matched to their intervention counterparts for final effectiveness and cost-benefit analyses. Participants enrolled in the comparison group are assigned a study identification number.

An embedded mixed method process analysis is being conducted to identify and mitigate challenges to implementation. Quantitative data are extracted from Epic and reviewed quarterly to inform reach and adoption. Quarterly data are used to adapt questions to further understand the quantitative data using qualitative methods. Qualitative data are subsequently analyzed using a narrative realist analytic approach. Meeting minutes, observation notes, and lifestyle coach feedback are compiled and analyzed in Word (Microsoft Corp) documents twice annually. Solutions to challenges are framed in the context of what is acceptable and appropriate to patients and providers and what is feasible within the context of the health care system (inner setting) and outer setting conditions, policies, and funding [33]. Findings are used in the ongoing planning process to continuously adapt, implement, and evaluate. Similar complex mixed methods designs have been used for evaluating the implementation of lifestyle interventions within new populations and settings [34-38].

Arithmetic means will be calculated overall and for specific sectors of health care between the 2 groups. Compound annual growth rates in per capita expenditures will be calculated to track the geometric progression ratio and estimate the rate of growth in spending time. A propensity score matching method will be implemented to compare case subjects to control subjects. The model takes the form of P(X)=Pr (D=1jX), where D=1 indicates DPP participation and matches on the probability of participation instead of attempting to create a match for each participant with the same value of X. This model aims to account for unobservables and obtain an unbiased and accurate measure of costs. Matched control subjects who do not participate in a DPP must be similar to the case subjects who did participate such that the only difference is attributed to the disease. The matching model in this analysis will control for the following: patient age; sex; race; ethnicity; zip code; and prevalence of other conditions such as but not limited to hypertension, chronic obstructive pulmonary disease, congestive heart failure, cancer, and metformin use in all years of the data. Differences in spending will be compared between individuals with prediabetes who participate in a DPP and those individuals who do not compute net savings and ROI (if any) of program participation. A similar methodology may be applied to compare individuals who progress to diabetes relative to those who do not progress for more in-depth cost savings and ROI analysis. Given this is the first time the DPP will be implemented at the UMMC, the sample size will be determined based on the convenience sample or the number of participants enrolled based on the proposed recruitment and retention strategies.