Abstract
Purpose:
Type 2 diabetes is a global pandemic, with 1 in 6 people expected to be diagnosed by 2050. The Diabetes Prevention Program (DPP) is an evidence-based program that has been shown to reduce A1c and bolster health outcomes in people with type 2 diabetes and prediabetes, but implementation has been varied, with lower uptake in rural and economically underserved communities. The study assessed whether there are geographic and socioeconomic disparities in the availability of hospital-supported DPPs in the US.
Methods:
We assessed DPPs in 3204 general medical hospitals. Data on hospital and county characteristics came from the American Hospital Association (AHA) Annual Survey, the Area Health Resource File, and County Health Rankings. We assessed geographic and socioeconomic differences between hospitals with and without DPPs using t-tests and chi-square tests. Rurality was determined based on the 2013 Rural-Urban Continuum Codes (RUCC). We then conducted a multivariable analysis to assess the relationship between geographic location, socioeconomic characteristics and the presence of a DPP, independent of hospital factors.
Findings:
Nearly half (49.3%, n = 1580) of hospitals in the U.S. reported having a DPP in 2021. DPPs were less commonly found in rural counties as well as micropolitan counties when compared to their urban counterparts. After controlling for hospital size and other community characteristics, these disparities disappeared. When including the interaction of hospital size and geographic location, the odds of offering DPPs was higher among large, rural hospitals. DPPs were also less commonly available in counties with more limited food access and in health professions shortage areas.
Conclusions:
DPP implementation is less common in rural and underserved areas. The interaction between hospital size and location is helpful for understanding barriers to DPP availability.
Introduction
Type 2 diabetes is a global pandemic, with 540 million people diagnosed. 1 In the United States, 38.4 million people currently live with diabetes and 1 in 4 healthcare dollars are spent managing diabetes. 2 Prevention of type 2 diabetes is critical. Evidence-based diabetes prevention programs exist such as the landmark Diabetes Prevention Program (DPP). 3 The DPP is a structured lifestyle intervention that is provided by a trained expert and is focused on dietary changes, physical activity, behavioral changes, group support, regular monitoring and long-term maintenance. 4 The DPP has demonstrated that lifestyle interventions regarding nutrition and physical activity can significantly reduce the risk of developing type 2 diabetes in people with prediabetes and has formed the basis for many prevention programs worldwide. While effectiveness has shown to be lower in racial and ethnic minority groups,5 -7 the DPP has demonstrated that lifestyle changes regarding nutrition and physical activity can significantly reduce the risk of developing type 2 diabetes in people with prediabetes. Despite the promise of the DPP, efforts to scale up this intervention have been variable, with reduced DPP availability in rural communities. 8
The DPP landmark longitudinal clinical study found that an intensive lifestyle intervention—focused on achieving 7% weight loss and 150 min of weekly physical activity—led to a 58% reduction in diabetes incidence. 9 One study group received Metformin while the second study group received intensive nutrition and physical intervention. Follow-up studies, including the Diabetes Prevention Program Outcomes Study, confirmed that lifestyle modifications provided long-term benefits, with a 34% lower diabetes risk even after a decade.3,4 These findings shaped public health initiatives like the CDC’s DPP and influenced Medicare coverage for DPP. Since the COVID-19 pandemic, DPPs are offered in person and online.
Many studies have sought to better understand the relationship between social determinants of health and the DPP but none have investigated the impact of rural and urban locations. Notably, urban and rural hospitals differ significantly in healthcare delivery due to variations in resources, workforce availability, and policy differences. Rural hospitals often struggle with limited specialty care access, longer patient travel distances, and persistent physician shortages, with less than 8% of doctors choosing to practice in rural areas. Importantly, rural hospitals are more reliant on Medicaid funding, making them particularly vulnerable to policy changes and financial instability. Roughly 1 in 5 people in the U.S. live in rural areas 10 and individuals who live in rural locations are more likely to experience type 2 diabetes, cardiovascular disease, and diabetes complications. 11 Efforts to improve rural health and rural health equity focus on enhancing access to care, recruiting and retaining healthcare professionals, promoting preventive services, addressing health disparities, leveraging technology, and strengthening community-based healthcare initiatives. 12
A study on the nationwide rollout of the DPP suggested that when hospitals partnered to support DPP implementation through screening, recruitment, and targeted referrals, participation in DPP improved. 13 Rural communities may be at a disadvantage for hospital-supported DPP, however, due to: (1) having fewer healthcare facilities, an ongoing health professions shortage, and limited transportation for completing the DPP 14 ; (2) rural hospitals may also be smaller, have fewer resources, and be located in communities with limited partner organizations to effectively screen, recruit, and support DPPs. In this study, we assessed the number of hospitals providing DPP to determine if rural-urban disparities in hospital-provided DPP exist and to determine additional hospital and community characteristics associated with DPP.
Methods
Research Design
Data for this secondary data analysis are from 3 data sources: the American Hospital Association (AHA) Annual Survey, the Area Health Resource File (AHRF), and County Health Rankings. While the AHA survey provides hospital-specific data, the AHRF captures broader healthcare infrastructure trends, and the County Health Rankings dataset emphasizes community health and social determinants, making them valuable for research on healthcare access, disparities, and system performance. Our primary research question was whether DPP availability varied by whether a hospital was located in a rural or urban county. We controlled for social determinants of health in the surrounding county given evidence that social factors shape access to the DPP.15,16 We also controlled for a robust group of hospital characteristics given longstanding evidence that hospital size and other organizational factors are associated with service offerings.17,18
Participants
The American Hospital Association annually conducts a survey of the approximately 6000 hospitals across the United States. Hospital data for this study was sourced from the 2021 AHA Annual Survey, the most current at the time of the study. After excluding specialty hospitals and federally owned facilities who are a small portion of the U.S. hospitals and are less likely to provide diabetes programming, the sample consisted of responses from 4271 general medical hospitals across 50 states (including public, nonprofit, and for-profit); however, 1062 hospitals did not respond to the key dependent variable question on diabetes prevention programming. After merging community-level variables from the 2019 County Health Rankings dataset (selected in order to precede the 2021 AHA data), 5 additional hospitals were removed due to missing data. The final analytic sample for this study consisted of 3204 general medical hospitals.
Data Collection
The AHA survey gathers hospital-level data through direct responses from hospitals, administrative records, and external validation, providing insights into hospital characteristics and their offered services. The AHRF compiles data from over 50 sources, including federal agencies (eg, HRSA, CMS, and U.S. Census), offering comprehensive information on health workforce distribution, hospital infrastructure, and population demographics at multiple geographic levels. The County Health Rankings data aggregates secondary data from national and state sources such as the CDC, U.S. Census, and BRFSS, focusing on county-level health outcomes, behaviors, and disparities.
AHA Survey
The key dependent variable for this study is a binary (yes/no) variable of whether a hospital offers DPP, defined by AHA as a: “Program to prevent or delay the onset of type 2 diabetes by offering evidence-based lifestyle changes” 11 and sourced from the 2021 AHA Annual Survey. To determine additional hospital and community characteristics associated with DPP, the analysis also considered size, ownership, teaching hospital status, system membership, and religious affiliation, all sourced from AHA data, as well as geographic region.
County Health Rankings
From the County Health Rankings data, we incorporated county characteristics. These included income inequality in the county, measured as the ratio of household incomes at the 80th income percentile to those at the 20th income percentile; the prevalence of diabetes for the population, measured as the percentage of adults aged 20 years and above with diabetes; and food environment index, defined by County Health Rankings as accounting for “both proximity to healthy foods and income.” 15 This index is measured from 1 to 10 with 10 being the best food environment. Two additional county characteristics, the uninsured rate and whether the county is classified as a health professional shortage area (HPSA) were sourced from the AHRF.
RUCC
We included a measure of rurality, based on the 2013 Rural-Urban Continuum Codes (RUCC). For bivariate analyses, we consolidated the 9 RUCC codes into 3 classifications: metropolitan/urban (codes 1-3 on the continuum, consisting of counties in metropolitan areas); micropolitan/suburban (codes 4-6 on the continuum; consisting of counties adjacent to metro areas and/or with urban populations of greater than 20 000); and rural (codes 7-9 on the continuum, with completely rural populations or urban populations of less than 20 000) to better understand geographic subgroups. For multivariable analyses, we retained the continuous variable to facilitate interpretation of the interaction term.
Analysis Methods
The analytic plan consisted of producing descriptive statistics for each variable, both for the overall sample and for rural, urban, and micropolitan subgroups. We utilized t-tests and χ² tests to understand the differences between these groups. We then conducted a stepwise, logistic regression analysis due to the binary nature of our dependent variable. In this approach, we assessed the relationship between the presence of a DPP in (1) an unadjusted model, (2) after only adjusting for community characteristics (3) after adjusting for both hospital and community characteristics, and (4) after including an interaction term between hospital size and the urban/rural continuum. To further examine the nature of the interaction, we estimated the relationship between geographic location and bed size at 1 standard deviation below the mean, the mean, and 1 standard deviation above the mean for bed size.
Results
Descriptive Results
Nearly half (49.3%, n = 1578) of hospitals in the study sample reported having a DPP in 2021 (Table 1). Figure 1 depicts the number of DPPs per 10 000 residents by county, based on this national sample. Importantly, Figure 1 also shows a visual representation of the DPP deficit in Appalachia. As shown in Table 1, DPPs were underrepresented in in rural counties with 14.6% of DPP programs existing in rural counties, while rural hospitals represented 16.39% of the full sample (P = .006). DPP programs were also underrepresented in micropolitan hospitals (21.5%, P = .018), while they were overrepresented in urban hospitals (63.9%). When looking at prevalence of DPP programs within county types, 43.8% of rural hospitals were identified as having a DPP, compared to 45.4% of micropolitan and 52.2% of urban (see Table 2). Geographically, DPPs were overrepresented among Midwest and Northeast hospitals, 54.5% of Midwest hospitals and 62.8% of Northeast hospitals within the sample offering programs. However, DPP were underrepresented in the South, at only 40.4% of hospitals.
Hospital Characteristics by Diabetes Prevention Program Status, n = 3204.
Abbreviation: SD, standard deviation.
P-values calculated using 2-sample t-tests for continuous variables and chi-square tests for categorical variables.
Hospital diabetes prevention programs per 10 000 county residents.
Diabetes Prevention Program Prevalence by County and Hospital Characteristics.
Note. DPP, Diabetes Prevention Program.
Percentages in each row add to 100%. P-values from chi-square tests comparing DPP availability across categories.
DPPs were also less common in smaller hospitals (P ≤ .001) and in hospitals with a religious affiliation. Hospitals that were overrepresented in the subgroup with DPPs included those with major teaching status (10.65%, compared to 6.6% in the full sample, P ≤ .001) and nonprofit hospitals (75.4%, compared to 68.6% in the full sample; P ≤ .001). At the community level, DPPs were underrepresented in communities with poor food access (P ≤ 0.001), high diabetes prevalence (P ≤ 0.001), in health professional shortage areas (P ≤ .001), and in the South (P ≤ .001).
Regression Results
Using logistic regression, we found significant geographic differences in the availability of DPP (Table 3). After controlling community characteristics, rural hospitals had lower odds of offering the DPP (OR = 0.96). After adjusting for both community and hospital characteristics, population density was no longer a significant predictor of hospitals offering the DPP. To further probe the relationship between institutional and geographic resources, we added an interaction term between county-level density and hospital bed size in a second multivariable model. This interaction was significant, and we found that the odds of a hospital having a DPP significantly increased as bed size and rurality increased. In other words, these programs are especially common in larger rural hospitals. In plotting the interaction, we found small urban and rural hospitals have similar odds of having a DPP, but the effect of bed size becomes particularly pronounced in hospitals above the mean bed size (Figure 2).
Stepwise Multivariable Logistic Regression Predicting Hospital Diabetes Prevention Program Adoption, n = 3204.
Abbreviations: OR, odds ratio; SE, standard error.
Model 1: Base model with county diabetes rate
Model 2: Adds food environment and regional controls
Model 3: Adds hospital characteristics
Model 4: Adds interaction between county diabetes rate and hospital beds
P < .05. **P < .01. ***P < .001.
DPP availability and total hospital beds.a
Discussion
Fewer rural and micropolitan hospitals offer the DPP in the U.S. as compared to their urban hospital counterparts. The fact that DPPs are less common in rural hospitals compared to their urban counterparts has important implications given the disproportionate impact type 2 diabetes has in rural areas. 14 These differences disappeared after controlling for organizational and community characteristics suggesting that other factors, such as hospital size and the level of community resources, may be associated with both rural location and DPP implementation. Importantly, we found a significant interaction between rurality and hospital bed size, such that larger rural hospitals had increased odds of implementing a DPP. Our findings show that DPP was especially common in larger rural hospitals and to further understand that relationship, we plotted the interaction of bed size. We found small urban and rural hospitals have similar DPP odds and hospitals with more beds are more likely to offer DPP. Smaller hospitals are challenged by financial issues, critical staffing shortages, and higher fixed costs per patient. 15 This finding suggests that larger, rural hospitals may have additional resources to successfully support a DPP in their community.
We also found that several other community characteristics were associated with whether hospitals supported a DPP in their community which have important implications for equity in diabetes care. For example, hospitals located in counties with high diabetes prevalence and higher uninsurance rates were less likely to have a DPP. Hospitals located in counties with strong food access had higher odds of supporting a DPP, as did hospitals located in areas not designated as a HPSA. These findings suggest that the reasons rural hospitals are less likely to offer DPP is potentially because these communities also disproportionately have few resources and greater diabetes burden. 16 For example, in areas with HPSAs, there is a lack of trained professionals to provide supportive diabetes prevention services such as health communication, and nutrition counseling that are essential components of the DPP. Similarly, offering the DPP requires access to fresh foods which may mean that hospitals are less able to support this intervention when located in poor food environments.
These findings have important implications for public health given that reduced access to DPPs in rural areas may exacerbate health outcomes secondary to type 2 diabetes. Previous studies have demonstrated numerous barriers to implementing and sustaining the DPP, including financial resources, social determinants of health, and program fidelity. 17 Although partnerships with hospitals are an important facilitator to successful DPPs in previous studies, 12 rural hospitals may be less able to provide this support. Our study suggests that tailored implementation strategies for rural hospitals may be necessary to overcome a dearth of resources and staffing shortages in these areas. Previous research suggests utilizing certified diabetes care and education specialists and integrating screening and referrals into clinical workflows are important strategies for supporting implementation of DPP. 18 Partnering with community organizations and public health departments, while seeking sustainable funding and engaging trusted local voices, can also enhance reach and long-term success. 19 These strategies may be particularly important in rural communities where social factors such as food and health care access limit uptake of evidence-based programs such as the DPP.
Limitations
The main limitation of this study is its reliance on AHA data. While this data is robust in many ways, it is self-reported in nature, meaning that none of the diabetes programming reported has been independently verified. Additionally, the survey question regarding diabetes programming is a simple yes/no response. This means we are lacking any data on variation between hospital programs, a hospital’s level of investment in a program, or other nuance that would be helpful in a better understanding of these efforts. Given that responses to this question are based upon a hospital’s interpretation of the definition, there is potential that variation in how DPP is defined may have substantial impact on program outcomes. DPP could be overreported. A related limitation is that there are many types of educational programs for diabetes prevention that are not tracked in the AHA annual survey. Future research should explore adoption of different forms of diabetes prevention within hospitals. Finally, we recognize that approximately 1 quarter of the hospitals that met our inclusion criteria did not respond to the survey question relevant to our key dependent variable. To better understand the characteristics of those that did not respond, we compared the respondents to nonrespondents and found that these hospitals were smaller, less likely to be an academic medical center, and less likely to be part of a hospital system. We are unable to determine whether these hospitals simply did not complete this section of the survey or whether they opted out of the question because of a lack of services in this area.
Conclusions
DPP is less common in rural and otherwise underserved areas when measuring food security status and HPSAs. Importantly, these programs are also less common in under-resourced institutions that are likely experiencing rural health disparities. Given the societal health impact of diabetes as a chronic disease, policymakers should consider avenues for better supporting interventions in under-resourced areas, specifically focusing on rural location and hospital size.
Footnotes
Acknowledgements
None.
Ethical Considerations
There are no human participants in this article and informed consent was not required. The authors declare no ethical issues. All methods were carried out in accordance with relevant guidelines and regulations.
Consent to Participate
The data used was from a national database so informed consent was not needed.
Consent for Publication
Not applicable.
Author Contributions
Allyson Hughes: Conceptualization, Writing, Methodology, and Data interpretation. Berkeley Franz and Cory Cronin: Data curation and Writing. Shyamkumar Srirarm: Writing—Reviewing and Editing.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
The datasets generated and analyzed in this study are available from the corresponding author upon reasonable request.