Social Vulnerability and National Diabetes Prevention Program Recognition Status

Abstract

Introduction:

The CDC National Diabetes Prevention Program (National DPP) aims to reduce the incidence of type 2 diabetes in the U.S. Organizations delivering the National DPP receive pending, preliminary, full, or full-plus recognition status based on specific program criteria and outcomes. Achieving full/full-plus recognition is critical for organizations to sustain the program and receive reimbursements to cover costs, but organizations in disadvantaged areas may face barriers to obtaining this level of recognition. This study examined the association between county-level social vulnerability and full/full-plus recognition status within the National DPP.

Methods:

Using the 2022 National DPP registry and the 2018 CDC Social Vulnerability Index (SVI), a three-level categorical dependent variable was created (n=843): counties without organizations having full/full-plus recognition, counties with at least one organization not having full/full-plus recognition, and counties with all organizations having full/full-plus recognition. A multinomial logit model was analyzed in 2023 to examine the association between SVI and in-person full/full-plus recognition organizations at the county level, adjusting for confounders.

Results:

Compared to counties with low social vulnerability, counties with higher social vulnerability had significantly higher odds of having no organizations with full/full-plus recognition. For example, counties with high SVI had 2.63 (95% CI: 1.55–4.47) times higher odds of having no organizations with full/full-plus recognition compared to having all organizations with full/full-plus CDC recognition.

Conclusions:

The findings suggest disparities in the National DPP recognition status among organizations in vulnerable communities. Developing strategies to ensure organizations in high social vulnerability areas achieve at least full recognition status is critical for program sustainability and reducing diabetes-related health disparities.

INTRODUCTION

The CDC National Diabetes Prevention Program (National DPP) is a nationwide effort to reduce type 2 diabetes incidence in the U.S. CDC partners with community and healthcare organizations to deliver this evidence-based program targeting weight loss through healthy lifestyle change.¹ To deliver the National DPP lifestyle change program (LCP), organizations must request, achieve, and maintain CDC recognition, classified as pending, preliminary, full, and full-plus recognition.² Full and full-plus recognition is granted to organizations that can demonstrate, for example, that 60% of participants-completers achieve a 5% weight loss or a 0.2% reduction in baseline HbA1c. More information can be found in the Diabetes Prevention Recognition Program Standards and Operating Procedures.² Full-plus recognition also requires meeting specific participant retention rates.² Organizations with full or full-plus recognition can request reimbursement through public payers and commercial health insurance plans, which represents a sustainable funding stream to support organizations delivering the program.^3,4 Concerns arise regarding the attainability of full or full plus recognition for organizations serving communities with low socioeconomic status due to challenges in recruiting/retaining participants and meeting required participant-level outcomes.^5,6 Not achieving full or full-plus recognition potentially threatens the ability of organizations located in disadvantaged communities to deliver the program, which could widen diabetes-related disparities.⁵ This study assessed the association between social vulnerability and full/full-plus recognition status among organizations delivering the in-person National DPP LCP at the county level.

METHODS

This study linked data from the 2022 National Registry of Recognized Diabetes Prevention Programs, a list of all CDC-recognized organizations,⁷ with the 2018 CDC Social Vulnerability Index (SVI) to determine county level social disadvantage.⁸ Covariates were obtained through Census data (Gazetteer and County Population by Characteristics files) and the CDC Diabetes Surveillance System file.^8–11 The study was exempted by the University of Central Florida IRB (STUDY00004824).

The National DPP Registry contained information about 2,218 organizations, including addresses and organizations’ recognition status (pending, preliminary, or full/full-plus).⁷ Only organizations offering the program in-person were included, and they were located in 843 counties. Other delivery modes (online, distance learning, and combination) were excluded as their addresses likely represent organizational offices rather than locations where the program was delivered. A three-level categorical dependent variable was created to classify counties based on organization recognition status: counties without organizations having full/full-plus recognition (none full); counties with ≥1 organization not having full/full-plus recognition (partially full); and counties with all organizations having full/full-plus recognition (all full). Hereafter full/full-plus is referred to as full. The exposure variable, SVI, was calculated using 15 social factors comprising four domains: socioeconomic status, household composition/disability, minority status/language, and housing type/transportation. Detailed information about these factors is available in the Appendix and has been described elsewhere.⁸ Higher values indicate greater social vulnerability, which the CDC categorizes as Low, Low-Medium, Medium-High, and High at the county level.⁸ Four covariates were created: population density per square mile (<250 vs. ≥250), proportion of females (<51% vs. ≥51%), proportion of adults with diabetes (<10% vs. ≥10%), and whether organizations in the county were open to the public (none vs. at least one). All covariates were binary variables; cutoffs were determined based on data distributions and/or national averages.

Chi-squared tests were used to examine basic associations between full recognition status and SVI, as well as the covariates. A multinomial logit model was used to examine the association between SVI and full recognition status at the county level, adjusting for covariates. All analyses were conducted using SAS Enterprise Guide 7.1, with p<0.05 indicating statistical significance.

RESULTS

Table 1 presents the characteristics of the 843 counties with organizations delivering the in-person National DPP LCP, categorized by recognition status (none full, partially full, and all full). A total of 38% of counties had no organizations with full recognition, 22% had at least one organization without full recognition, and 39% had all organizations with full recognition. Among counties without organizations achieving full recognition, 26% had a high SVI, compared to 23% among counties having at least one organization without full recognition and 16% among counties where all organizations have achieved full recognition.

Table 1.

Characteristics of Counties With Organizations Delivering the National DPP by Full CDC Recognition

Variable	Total	None full	Partially full	All full	p-value
Overall, n (%)	843	324 (38.4)	190 (22.6)	329 (39.0)
Social Vulnerability Index (SVI), n (%)					<0.001
Low	188	57 (17.8)	31 (16.4)	100 (30.5)
Low-medium	226	94 (29.3)	49 (25.9)	83 (25.3)
Medium-high	242	85 (26.5)	65 (34.4)	92 (28.1)
High	182	85 (26.5)	44 (23.3)	53 (16.2)
Percentage of the population with diagnosed diabetes, n (%)					0.016
<10%	597	230 (71.0)	120 (63.2)	247 (75.1)
≥10%	246	94 (29.0)	70 (36.8)	82 (24.9)
Population density per square mile, n (%)					<0.001
<250	578	266 (82.1)	64 (33.7)	248 (75.4)
≥250	265	58 (17.9)	126 (66.3)	81 (24.6)
Percentage of females in the county, n (%)					<0.001
<51% females	356	144 (44.4)	49 (25.8)	163 (49.5)
≥51% females	487	180 (55.6)	141 (74.2)	166 (50.5)
Whether organizations in the county were open to the public, n (%)					<0.001
None	54	36 (11.1)	0 (0.0)	18 (5.5)
At least one	789	288 (88.9)	190 (100.0)	311 (94.5)

Note: Boldface indicates statistical significance (p<0.05).

^a:Where the total does not add up to 843, counties had missing data for that characteristic;

^b:Results are based on bivariate analyses using chi-squared tests comparing three levels of counties with full CDC recognition to deliver the National DPP across levels of county characteristics;

^c:Social Vulnerability Index from the CDC/ATSDR.

Table 2 presents the fully adjusted results of the multinomial logistic regression model. In general, the odds of having recognized organizations decreased progressively as SVI increased, demonstrating a graded association. Compared to counties with low SVI, counties with higher SVI had significantly higher odds of having no organizations with full recognition or at least one organization without full recognition than having all organizations with full recognition. For instance, counties with low-medium SVI had 1.95 (95% CI: 1.25–3.04, p=0.003) times higher odds of having no organizations with full recognition compared to having all organizations with full recognition. Similarly, counties with medium-high SVI had 1.61 (95% CI: 1.02–2.55, p=0.041) times higher odds of having no organizations with full recognition compared to having all organizations with full recognition. Finally, counties with high SVI had 2.63 (95% CI: 1.55–4.47, p<0.001) times higher odds of having no organizations with full recognition compared to having all organizations with full CDC recognition.

Table 2.

Multinomial Logistic Regression of County Characteristics Associated With National DPP Organizations by Full CDC Recognition

Social Vulnerability Index (SVI)	None full vs. all full OR (95% CI)	p-value	Partially full vs. all full OR (95% CI)	p-value
Low (reference group)
Low-medium	1.95 (1.25–3.04)	0.003	1.83 (1.03–3.28)	0.039
Medium-high	1.61 (1.02–2.55)	0.041	2.06 (1.15–3.71)	0.015
High	2.63 (1.55–4.47)	<0.001	3.03 (1.53–6.02)	0.001

Note: Boldface indicates statistical significance (p<0.05).

^a:None full = Counties without organizations having full recognition, Partially full = Counties with ≥1 organization not having full recognition, All full = Counties with all organizations having full recognition;

^b:Odds ratio adjusted for the percentage of the population with diagnosed diabetes, population density, percentage of female individuals in the county, and whether the program was offered to the general public (Table 1).

DISCUSSION

This study examined the association between county level social vulnerability and full CDC recognition status among organizations delivering the National DPP LCP. The findings indicate that counties with higher social vulnerability exhibited greater odds of lacking fully recognized organizations. Without full recognition, community and healthcare organizations may encounter greater challenges in maintaining program continuity due to potentially reduced stability and lack of financial reimbursement.

These findings expand upon previous research demonstrating that counties characterized by heightened diabetes risk and socioeconomic disadvantage had fewer available National DPP organizations.¹² Through a health equity lens, this study emphasizes the importance of addressing community disadvantage for the ongoing implementation and sustainability of the National DPP. Future research should investigate factors hindering the achievement and maintenance of full CDC recognition. To ensure equitable access to this program, additional support could be provided to organizations operating in areas of heightened social disadvantage to achieve and maintain full recognition. Adapting the performance-based requirements for full CDC recognition may be necessary to overcome unique challenges faced by National DPP-delivering organizations located in socially vulnerable communities. Sustainable funding models (e.g., risk-adjusted payment models) that account for community-level characteristics using the SVI or similar indices should be considered. Additional strategies could include implementing the National DPP as part of Medicaid coverage in all states, improving reimbursement rates, streamlining recognition processes, and providing administrative billing support.

Limitations

The study has several limitations. Associations observed at the county level may not accurately represent associations at the individual level. Further, there may be other factors that influence the association between social vulnerability and recognition status. For example, although the SVI utilizes various community-level factors and the analysis was adjusted for confounders such as population density, other community characteristics, resources, and infrastructure—particularly in larger counties with diverse populations and systems—may have impacted the results. Moreover, it is important to recognize that other LCPs not part of the National DPP may exist within communities. Additionally, recognition status, though designed to ensure the fidelity and quality of programs nationwide, does not necessarily reflect the program’s effectiveness. A strength of this study is the use of a publicly available nationwide dataset of organizations delivering the program.

CONCLUSIONS

In conclusion, disparities in the CDC National DPP recognition status were observed among organizations in vulnerable communities, raising awareness of the need to develop strategies to overcome potential barriers for these organizations, as mentioned above. Policymakers can use the findings from this study to initiate further research and policy development, prioritizing the needs of the most vulnerable communities where existing diabetes-related health disparities are frequently documented.