Patient-centered Goal Setting in the National Diabetes Prevention Program

Diabetes affects 13.0% of US adults, a rising trend that especially impacts racial and ethnic minority and low-income individuals. In response, the Centers for Disease Control and Prevention launched the National Diabetes Prevention Program (NDPP) in 2010 to provide evidence-based lifestyle intervention for adults with risk factors such as overweight/obesity and prediabetes (e.g., A1C 5.7-6.4%). Successes include reaching >300,000 participants and delivery in all 50 states. However, a major concern is limited effectiveness for disadvantaged groups with the highest risk of diabetes. For example, Latinx are the largest ethnic minority group in the US, but least likely to enroll in or complete the yearlong NDPP, which may widen disparities and diminish overall program impact. The NDPP focuses primarily on weight loss to reduce risk, but racial/ethnic minority and low-income groups both lose about half as much weight as white and higher-income groups, respectively (~2% vs. ~4%).

To address these challenges, we adapted the NDPP with an array of personalized goals for risk-reduction to meet individual needs (NDPP-Flex). We developed NDPP-Flex based on the Health Belief Model in which factors like personal characteristics, perceived barriers, and self-efficacy influence health outcomes. We also focused on no-cost modifications per our extensive prior research on NDPP implementation that identified both strengths (e.g., cost-effectiveness) and weaknesses (e.g., low reimbursement rates). We piloted NDPP-Flex with 95 diverse participants (e.g., 76% Latinx) in Federally Qualified Health Centers (FQHCs). We compared outcomes to a demographically matched sample (n=245) from our prior delivery of the standard NDPP with pre-set goals for ≥150 minutes of physical activity per week and ≥5% weight loss. NDPP-Flex participants had similar attendance, physical activity, and weight loss, yet achieved about four times greater A1C improvement (0.22 ± 0.05% vs. 0.06 ± 0.03%; p=.018) and likelihood of normoglycemia at follow-up (OR 4.62; p=.013 [95% CI 1.38-15.50]). Clinically, these outcomes are highly significant as a 0.2% A1C reduction is associated with 58% lower risk of diabetes at 3 years, and a return to normoglycemia predicts 56% lower long-term incidence of diabetes. Our findings were published in Diabetes Care along with an invited commentary which noted that our approach is "innovative" and "should be adopted as part of our national strategy of health promotion."

We propose a cluster-randomized controlled trial of NDPP-Flex versus the standard NDPP to evaluate effects on glycemia, heterogeneity of treatment effects, and mechanisms of change. We will recruit 200 diverse adults with prediabetes (A1C 5.7-6.4%) and normal weight or overweight/obesity. Participants will enroll in classes (N=20) that are randomized to deliver NDPP-Flex or the standard NDPP. The trial is designed to have >95% power to detect anticipated treatment effects of ~0.2% A1C improvement and ~10% increase in participants who reach normoglycemia compared to the control arm. The study will be conducted by an expert team in a health system with the 8th largest network of FQHCs in the US (Denver Health) and in partnership with a premier research institution (University of Colorado). Specific aims are:

Aim 1. To assess effects of NDPP-Flex on A1C change (primary outcome) and achieving normoglycemia (A1C <5.7%) at 12-months (secondary outcome). Hypothesis: Compared to the standard NDPP, NDPP-Flex will reduce A1C and increase the likelihood of normoglycemia.

Aim 2. To assess heterogeneity of treatment effects on A1C by demographic characteristics (age, gender, race/ethnicity, income, language) and for individuals of normal weight (vs. overweight/obesity) to understand for whom NDPP-Flex may be effective. Hypothesis: For all subgroups of participants, NDPP-Flex will result in equivalent or superior A1C improvement than the standard NDPP.

Aim 3. To explore effects of NDPP-Flex on other health outcomes (weight, diet, physical activity) and psychosocial outcomes (motivation, perceived control, perceived stress). We will explore these outcomes as mediators of glycemic improvement to understand the key mechanisms of change. Hypothesis: NDPP-Flex will have a greater positive effect on psychosocial factors than the standard NDPP, which will account for greater A1C improvement than differences in health outcomes.

Impact: NDPP-Flex could shift clinical guidelines from prescriptive (and largely unattainable) goals to a more person-centered approach to diabetes prevention. Moreover, NDPP-Flex will be preferable from a health equity lens if it is better for underserved groups and without diminishing benefits for others. This approach could also expand public health impact by reaching lean adults with prediabetes who are currently excluded from the NDPP (given its focus on overweight/obesity), but account for ~10-20% of incident diabetes cases. The results have strong implementation potential through rapid dissemination among >2000 organizations that currently deliver the NDPP, and by contributing high-quality evidence to modernize clinical guidelines for prevention.