The role of Sociodemographic factors on goal achievement in a community-based diabetes prevention program behavioral lifestyle intervention

An estimated 10.5% of the US adult population has diabetes, the majority of which is type 2 diabetes [1]. Various estimates demonstrate disproportionately higher diabetes prevalence in lower income and education groups [2, 3] and in non-Hispanic black individuals compared to non-Hispanic white (16.4% vs. 11.9% [1]). Diabetes is associated with complications such as cardiovascular disease, end-stage renal disease, and premature mortality, [4] which are also disproportionately prevalent among many of these same sociodemographic groups [5, 6] who are likely disadvantaged due to lower income and education, discrimination, and related social factors [3]. Thus, approaches to diabetes prevention that are effective across sociodemographic subgroups are a priority.

The landmark Diabetes Prevention Program (DPP) demonstrated that a behavioral lifestyle intervention was highly effective in decreasing or delaying the onset of type 2 diabetes among a diverse cohort of overweight/obese individuals with prediabetes [7]. Due to the success of translated versions of the DPP lifestyle intervention, the US Centers for Disease Control and Prevention (CDC) established a system for widescale dissemination of the intervention program [8]. Relatively recently, the Centers for Medicare and Medicaid Services (CMS) began offering reimbursement for CDC approved DPP-based lifestyle intervention [9]. With widespread implementation of DPP-based lifestyle intervention programs, it is important to ensure reimbursable programs are effective among sociodemographic groups who carry a disproportionate burden of type 2 diabetes.

The DPP-Group Lifestyle Balance (DPP-GLB), one of the CDC recognized/CMS reimbursed DPP-based lifestyle interventions, has shown to be effective in a variety of diverse community settings [10‐15]. However, identifying potential disparities in uptake and maintenance of recommended behavioral changes during the course of the DPP-GLB would be beneficial in directing efforts to ensure equitable intervention effectiveness. It is also worth considering the potential for the DPP-GLB intervention to address disparities in health behaviors that may have existed at intervention baseline [3]. Identifying potential disparities in achieving the primary behavioral goals of the DPP-GLB could inform the need to modify the current intervention with possible solutions that include additional support and resources to promote success among disadvantaged sociodemographic groups. A better understanding of disparities in lifestyle intervention progress could also speak to the upstream influence of social determinants of health [16].

The primary goals of the DPP-GLB are the same as the DPP itself, to achieve a 7% weight loss (with a 5% weight loss accepted as the criteria for program success in these translation efforts; per CMS protocol [8, 9]), and to reach 150 min per week of moderate or greater intensity physical activity [7, 17]. We evaluated success in reaching these primary intervention goals across sociodemographic subgroups during a community-based DPP-GLB program.

For this current effort, we combined participant data from two clinical trials offering a community-based DPP-GLB intervention, the Healthy Lifestyle Project (GLB-Healthy, 2010–2014) [12] and The Physical Activity and Sedentary Behavior Change study (GLB-Moves, 2014–2019). Both studies received University of Pittsburgh Institutional Review Board approval and all subjects provided written informed consent.

A total of 240 participants (85%) had complete data available at baseline and after 6 and 12 months of intervention. As shown in Table 1, 74.2% of participants were women and the mean age of participants was 62.5 years. The majority of participants (93% of those included in race/ethnicity analyses) identified as NHW. Most participants were retired (42.5%), and all but 12.5% had completed at least some college education. Baseline weight and activity across sociodemographic factors are also shown in Table 1.

Tests for independence of sociodemographic variables found that race/ethnicity and income (p < .001), employment and education (p < .001), and income and education (p = 0.03) were not independent. All other relationships between sociodemographic variables did not deviate from independence. Overall median attendance was 21 [interquartile range (IQR) 17–22] sessions, out of a possible 22, with a median of 20 or higher for each sociodemographic subgroup.

As a whole, the DPP-GLB lifestyle intervention was effective in helping participants across most sociodemographic groups lose weight and meet their physical activity goals. However, despite widespread success due to intervention participation, disparities existed in the amount of weight lost between groups with, for example, NHW showing greater weight loss than NHB participants. We also found that retired participants had relatively greater weight loss than other employment groups at the end of the maintenance phase of the intervention. In addition, it appeared that the intervention may have begun to help equalize meeting the physical activity goal by 6 months among women, NHB, and those with lower income/education, all who had lower activity levels at baseline. Unfortunately, the disparities in activity levels found at baseline and attenuated at 6 months began to emerge again by 12 months. These findings identify the need for more effective, individualized approaches during the intervention maintenance phase such as offering additional support in order to address the complex influence of existing sociodemographic factors that impact the maintenance of weight loss and increased physical activity.

Our race/ethnicity weight loss results are in line with findings from the DPP efficacy trial and in DPP-based lifestyle intervention community translation efforts. In the DPP multicenter clinical trial, white relative to black participants had more success with weight loss after the 6 month core and at the end of the formal trial, after around 3 years [22]. In CDC recognized DPP-translation programs, NHB individuals were less likely to achieve a 5% weight loss compared to NHW [23]. Studies exploring weight loss success in DPP-translation efforts in black populations have also found suboptimal weight loss [14, 24, 25]. This consistent finding of less weight loss among NHB participants compared to NHW emphasizes the need to modify existing DPP-based lifestyle interventions to enhance effectiveness among NHB individuals.

In the DPP clinical trial, there were no overall gender differences in the odds of meeting the weight loss goal at either the end of the core intervention or at study end [22] although black women had significantly smaller weight loss compared to other racial/ethnic groups [26]. Unlike the DPP clinical trial, DPP-translation findings have found a higher odds of meeting a 5% weight loss goal among men compared to women [8]. However, similar to the DPP clinical trial, translation efforts have echoed smaller weight loss among black women [24]. In our current effort, men lost more weight, in kilograms, than women at 6 months although we had too small a sample size to look at goal success by race/ethnicity*gender.

Significant weight loss differences were found by employment and education in this effort. In the DPP multicenter clinical trial, weight loss success differed significantly by employment, with retired participants demonstrating the highest percentage who were successful after the full intervention [22]. Our findings also demonstrated a relatively high weight loss success rate among retired participants. In regard to education, which was significantly correlated with employment in this current effort, our study suggested differences by education at 12 months, with the most success being among the “high school or less” group. Notably, the majority (70%) of the high school education group also had a non-working employment status, with most retired, which may have contributed to their weight loss success. These findings are in contrast to those in a combined sample of CDC recognized DPP-translation efforts in which there were not significant differences in weight loss success by education [8].

Weight loss success by income in the DPP clinical trial and DPP translation literature is inconsistent. Participants from lower income groups showed the highest weight loss success in the DPP trial, [22] although translation efforts have indicated less weight loss among some low income groups [27]. In contrast to employment and education, our findings showed no weight loss differences by income.

While weight loss success has been evaluated in a number of DPP-translation efforts, success in physical activity is less frequently documented [28]. In the DPP itself, Hispanic American, Native American and Asian American individuals, retired participants, low-income participants, and men were most likely to meet the physical activity goal, respectively, after the full intervention [22]. Our findings demonstrated that fewer NHB participants appeared to have met the activity goal compared to NHW at 12 months. Unlike the DPP clinical trial, participants with higher incomes were most likely to meet activity goals in our translation findings. Similar to the DPP clinical trial, we also saw a higher percentage of men meeting the activity goal at baseline and at 12 months.

To our knowledge, we are the first study to evaluate sociodemographic disparities in physical activity in a DPP-translation effort. The lack of significant differences in physical activity by any sociodemographic category at 6 months and general improvement in the percentage meeting the activity goal during the course of the intervention indicate that the intervention was successful early on in promoting increased activity among those most in need of improvement. More research is warranted regarding how to sustain the improvements found at 6 months into the maintenance phase of the intervention and beyond.

The driving factors for differences in weight loss and physical activity across sociodemographic subgroups are likely multifaceted and complex. Social determinants of health that lead to disadvantage among certain sociodemographic groups have been indicated to increase risk for diabetes and related cardiometabolic conditions [29‐31]. These social determinants likely contributed to differences observed in physical activity at baseline in this effort, and may have contributed to differences in behavior change and change maintenance during the course of the DPP-GLB intervention. Ongoing research into the upstream factors contributing to disadvantage (e.g. exposure to adversity, food security, built environment, etc.) and additional efforts to address these factors are essential pieces of the puzzle in promoting equitable approaches to disease prevention [29, 31].

This effort has several strengths. We were able to evaluate success in achieving physical activity and weight loss goals in a community DPP-based lifestyle intervention program offered over a span of almost a decade. Participants provided overwhelmingly positive feedback of the intervention, as has been previously reported [12], and demonstrated excellent attendance, with no meaningful differences in attendance across sociodemographic groups. Engaging intervention participants in their local communities likely contributed to excellent program attendance. Close collaboration with community partners also likely contributed in part to the successful implementation of this community-based DPP-GLB program. Finally, the DPP-GLB also offers a CDC recognized curriculum that is intentionally adaptable to individual needs and preferences, making it an ideal program for use across a variety of diverse settings.

While every effort was made to include participants across the socioeconomic spectrum and among a variety of communities, this effort is limited by the amount of racial/ethnic diversity in the study sample, likely due to the lack of diversity in the greater Pittsburgh area. Also, the fact that physical activity level was not part of the eligibility criteria allowed for some participants to be relatively active at baseline. Finally, income was defined using neighborhood level estimates, which may not directly reflect individual level income. It is possible that additional neighborhood level factors may also influence success in lifestyle intervention programs.

The DPP-GLB is an appealing intervention option with the potential to improve risk factors for developing type 2 diabetes and related metabolic conditions. In fact, there have already been a variety of cultural adaptations of the DPP lifestyle intervention to suit a diverse array of populations and settings [32‐37]. In agreement with such efforts, our current findings suggest that NHB and non-retired participants would benefit from additional support in achieving and maintaining weight loss, while women, NHB, lower education and lower income participants would benefit from additional support in maintaining adequate activity levels. Alongside population level approaches to promote health equity, additional adaptations of DPP-based lifestyle intervention programs, particularly during the maintenance phase when intervention contact is less frequent, should be considered to minimize these disparities in weight loss and physical activity found at program end.