Geographic Disparities in Access to Broadband, Ambulance Services, and Health Care and Implications for Telehealth Authors: Celia Jewell, Heidi O’Connor, Thomas E. Griffin, Sarah M. Mah, Brianna L. Holston, Erin Troiano, Yvonne C. Jonk Published: 2026 ABSTRACT This cross-sectional study examines whether broadband services sufficiently support use of telehealth in rural areas lacking ambulance and health care services. Introduction In the US, rural residents experience greater health care needs yet have less access to timely health services compared with urban residents.1 Over the past decade, telehealth expanded significantly, bridging rural care gaps across the health care continuum (ie, preventive, primary and ambulatory, prehospital emergency, community paramedicine, acute and specialized, and postacute and rehabilitative).1,2,3 Yet, compared with urban residents, people living in rural areas face greater challenges in accessing both broadband and telehealth services.4,5,6,7 In many rural communities, insufficient broadband infrastructure results in low internet speeds and poor connectivity.5 These infrastructural challenges limit the potential for telehealth use, which requires reliable and sufficient internet bandwidth to support video conferencing and other remote medical services.4,5,8,9 To advance telecommunications capability, in March 2024, the Federal Communications Commission redefined sufficient broadband speed by increasing the threshold from a minimum of 25 megabits per second (Mbps) of download and 3 Mbps of upload (ie, 25/3 Mbps) to a minimum of 100/20 Mbps.10 As of June 2024, only 63 residential fixed connections per 100 households met this new criterion nationwide.11 Even in areas with sufficient broadband to support telehealth, issues with affordability, low digital literacy, inadequate computing devices, and availability of telehealth services often hinder telehealth adoption.7,12,13 A current knowledge gap is to what extent populated geographic locations, especially rural areas, lack adequate broadband needed for telehealth (hereafter referred to as broadband deserts [BBDs]) and how many of these locations also have limited health care infrastructure (prehospital ambulance services, hospitals, and clinics). Methods This population-based cross-sectional study built on previous work that identified areas lacking ambulance services (hereafter referred to as ambulance deserts [ADs])14 and areas lacking health care (hereafter referred to as health care deserts [HCDs]).15 Using ambulance location data from September 2021 to February 2022 that were obtained from 41 state emergency medical services offices, we defined ADs as populated census blocks (ie, blocks with ≥1 resident) with a geographic center outside a 25-minute drive from an ambulance station. Geographic units used by the US Census Bureau to collect and tabulate socioeconomic data range in size from individual census blocks to block groups, census tracts, and counties.16 HCDs were defined as populated census blocks with a geographic center outside a 30-minute drive from a hospital or clinic. Rural areas were defined as census block groups with more than 50% of the population classified as rural. The study was determined not to be human participant research by the University of Southern Maine institutional review board. Thus, informed consent was not required. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cross-sectional studies was followed. This study’s objectives were to document national, regional, state, and rural populations (number and percentage) living in BBDs, defined as areas unable to support telehealth due to broadband speeds of less than 100/20 Mbps; AD and HCD intersections with broadband speeds greater than 100/20 Mbps that could support telehealth; and BBD, AD, and HCD intersections where populations had limited access to adequate broadband to support telehealth, emergency services, and health services. The study’s other objective was to identify the extent to which households had the capacity to use telehealth services by estimating the number and percentage of households that owned a computing device with or without a broadband subscription. We used a geographic information system framework to identify BBDs within 41 states previously identified as having ADs.14 We used the Federal Communications Commission Broadband Data Collection June 2024 public-use data,17 obtained through the ArcGIS Living Atlas of the World,18,19 to identify populated census blocks with at least 1 broadband serviceable location (BSL). A BSL is “a business or residential location in the US at which mass-market fixed broadband internet access service is, or can be, installed.”20 The Broadband Data Collection layer in the ArcGIS Living Atlas categorizes each BSL into 1 of 3 service tiers defined by speed and latency: (1) served, defined as any service with low-latency fiber, cable, copper, or licensed terrestrial fixed wireless offering minimum speeds of 100/20 Mbps; (2) underserved, defined as any service with low-latency fiber, cable, copper, or licensed terrestrial fixed wireless offering speeds between 25/3 Mbps and 100/20 Mbps; and (3) unserved, defined as a service that is not low-latency or less than 25/3 Mbps and any service provided by geostationary satellite, nongeostationary satellite, unlicensed terrestrial fixed wireless, or other. Statistical Analysis To estimate the number of people in each census block who were underserved or unserved, the number of underserved or unserved BSLs were divided by the total number of BSLs per block to derive the percentage of underserved or unserved BSLs. The resulting percentages were multiplied by the total population in the census block to derive weighted estimates of the number of people who were underserved or unserved in each block. The public-use 2020 US Census Demographic and Housing Characteristics file (obtained through the ArcGIS Living Atlas)21 contains the percentage of rural and urban residents within each block group. The US Census Bureau classifies urban as all areas composed of a densely settled core of census blocks containing at least 2000 housing units or at least 5000 people; rural areas are defined as anything not urban.22 To create the BBD dataset, the block-level Broadband Data Collection data were aggregated to the block group level and joined to the 2020 US Census Demographic and Housing Characteristics file.21 If the percentage of people living in a block group was 50% or more rural, the block group was categorized as rural. The 2021 to 2022 AD and HCD datasets were then merged into the BBD dataset, and populations in BBDs, ADs, and HCDs were quantified at the national, regional, state, and rural levels. The identification of HCDs was based on the location of rural health clinics (RHCs), federally qualified health centers (FQHCs), and short-term acute care hospitals certified by the Centers for Medicare & Medicaid Services (CMS) obtained from the CMS 2021-2022 Provider of Service file.23 Many private health care facilities, primarily in urban areas, are not included in this file. Thus, the identification of HCDs and analyses of BBD, AD, and HCD intersections were limited to rural residents. The 2019-2023 American Community Survey public-use 5-year estimates24 obtained through the National Historical Geographic Information System25 were used to quantify populations that owned computing devices and had broadband subscriptions. Computing devices included desktops, laptops, smartphones, tablets, or other types of computers. High-speed broadband internet services included cellular data plans or cable, fiber optic, or digital subscriber lines installed within households. Data were analyzed November 2024 to July 2025. All analyses were conducted using ArcGIS Pro, versions 3.4 and 3.5 (Esri)26; SAS, version 9.4 (SAS Institute Inc)27; and Stata, version 17 (Stata Corp LLC).28 Missing values were recoded using related variables to ensure a complete dataset. Results BBDs Across the US and by Region Within the 41 states included in this study with a total population of 249.1 million, 11.9 million people (4.8%) lived in BBDs, of whom 10.5 million were rural residents (88.1%) (Figure 1). Counties with relatively high numbers and/or concentrations of people living in BBDs were mainly rural and located in the West, the Great Lakes region, northern Maine, the Appalachian states, and the South (Figure 2). Due to its high population density, the South (n = 123.8 million) had the greatest number of people living in BBDs (6.5 million), with 5.7 million rural residents (87.0%). The Midwest (n = 66.0 million) had the highest proportion (91.5%) of rural residents living in BBDs (3.0 million). The West (n = 23.9 million) had 1.4 million (86.2%) and the Northeast (n = 35.3 million) had 392 684 (86.8%) rural residents living in BBDs. Bar Graphs Showing the Number and Percentage of Individuals Living in Broadband Deserts (BBDs) by US Census Region and Rural Residence BBDs were defined as populated census blocks unable to support telehealth due to broadband speeds of less than 100/20 megabits per second. Data are presented for the total population of individuals living in BBDs in the 41 states included in the study (N = 11 868 032) and the rural subset (n = 10 455 665). All population estimates were derived from the US Census Bureau weighted population counts. Data sources: 2024 Federal Communications Commission Broadband Data Collection17 and 2020 US Census Demographic and Housing Characteristics file.21 Maps of the Number and Percentage of Individuals Living in Broadband Deserts (BBDs) by Urban vs Rural County of Residence in 41 States BBDs were defined as populated census blocks unable to support telehealth due to broadband speeds of less than 100/20 megabits per second. All population estimates were derived from the US Census Bureau weighted population counts. Data sources: 2024 Federal Communications Commission Broadband Data Collection17 and 2020 US Census Demographic and Housing Characteristics file.21 The West had the highest percentage of the population living in BBDs (6.9%) followed by the South (5.3%) and the Midwest (4.5%), while the Northeast had the lowest percentage (1.3%) (Figure 1). In terms of rural populations, the West had the highest percentage of rural residents living in BBDs (31.9%) followed by the South (19.1%) and Midwest (17.7%). Focusing on rural residents living in ADs and HCDs, patterns similar to those for BBDs emerged, with the South having the greatest number of people and the West having the greatest percentage of people living in each desert type (Figure 3). However, the West had both the greatest number and percentage (298 355 [6.7%]) of rural residents in a BBD, AD, and HCD intersection. Bar Graphs Showing the Number and Percentage of Rural Populations in Broadband, Health Care, and Ambulance Deserts by US Census Region in 41 States Data are based on a total rural population of 56 539 678 people. All population estimates were derived from the US Census Bureau weighted population counts. Data sources: 2024 Federal Communications Commission Broadband Data Collection,17 2020 US Census Demographic and Housing Characteristics file,21 2021-2022 Centers for Medicare & Medicaid Services Provider of Services file,23 and 2020-2021 Maine Rural Health Research Center Ambulance Location Data.15 AD indicates ambulance desert (>25-minute drive from ambulance station); BBD, broadband desert (broadband speeds <100/20 megabits per second); HCD, health care desert (>30-minute drive from health care facility). Intersection of ADs and HCDs With Broadband Access The greatest number of rural populations with potential access to broadband service yet diminished access to health services was on the West and East coasts; in northwestern Montana; in portions of Nevada, New Mexico, New York, Maine, New Hampshire, and Florida; and in the northern portions of Minnesota. The highest proportions were in the West, Midwest, portions of Maine, and Texas (eFigure 2 in Supplement 1). Within the 41 states, 946 278 rural residents lived in an AD and HCD intersection with sufficient broadband available to support telehealth (Figure 3). The South had 536 309 (9.5%), the West had 220 770 (15.5%), the Midwest had 99 255 (0.6%), and the Northeast had 89 944 (1.6%). Intersection of BBDs, ADs, and HCDs The largest concentrations (number and percentage) of rural populations living in BBD, AD, and HCD intersections were in the Western states, western Nebraska, northern Minnesota, and the northern parts of New England (eFigure 3 in Supplement 1). Within the 41 states, over 649 225 rural residents lived in a BBD, AD, and HCD intersection (Figure 3). The West had 298 355 (6.7%), the South had 251 316 (0.8%), the Midwest had 69 028 (0.4%), and the Northeast had 30 527 (0.5%). Computer Ownership and Broadband Subscription Rates by Region Within the 4 census regions, a mean (SE) of 96.7% (0.001%) of all households owned a computing device; rural households had slightly lower mean (SE) percentages (94.7% [0.006%] to 96.3% [0.009%]) than urban households (96.5% [0.003%] to 98.0% [0.003%]) (eFigure 1 in Supplement 1). The differences between overall and rural mean rates of access widened when taking into account households with a computing device and a broadband subscription. The overall mean (SE) percentage of households that had a computing device with broadband was 91.7% (0.002%). The mean (SE) rate in urban areas was 92.6% (0.002%), while in rural areas, it was 88.5% (0.004%). In urban areas, the mean (SE) rate ranged from 92.3% (0.003%) in the South to 93.9% (0.005%) in the West, and in rural areas, the mean (SE) rate ranged from 87.2% (0.006%) in the South to 92.1% (0.002%) in the Northeast (eTable 1 in Supplement 1). Rural BBDs by State The 10 states with the greatest number of rural residents living in BBDs consisted of 6 Southern states (Texas, North Carolina, Virginia, Alabama, Florida, and Georgia), 3 Midwestern states (Wisconsin, Missouri, and Michigan), and Washington in the West (Figure 4). The West had the highest percentage (31.9%) of rural populations living in BBDs. Six Western states had 29.1% to 39.3% of their rural populations living in BBDs (Wyoming, Washington, Montana, Idaho, Oregon, and New Mexico). Six Southern states had 20.1% to 32.2% of their rural populations living in BBDs (Louisiana, Florida, Virginia, Mississippi, Alabama, and Kentucky). Four Midwestern states had 20.2% to 27% of their rural populations living in BBDs (Wisconsin, Missouri, Illinois, and Nebraska) (Figure 5). Bar Graphs Showing Rural Populations in Broadband, Health Care, and Ambulance Deserts by State and US Census Region in 41 States Data are based on total rural population counts for each census region. States are ordered within regions by descending population count. All population estimates were derived from the US Census Bureau weighted population counts. Data sources: 2024 Federal Communications Commission Broadband Data Collection,17 2020 US Census Demographic and Housing Characteristics file,21 2021-2022 Centers for Medicare & Medicaid Services Provider of Services file,23 and 2020-2021 Maine Rural Health Research Center Ambulance Location Data.15 AD indicates ambulance desert (>25-minute drive from ambulance station); BBD, broadband desert (broadband speeds <100/20 megabits per second); HCD, health care desert (>30-minute drive from health care facility). Bar Graphs Showing Percentage of State Populations Living in Rural Broadband, Health Care, and Ambulance Deserts by US Census Region in 41 States All population estimates were derived from the US Census Bureau weighted population counts. Data sources: 2024 Federal Communications Commission Broadband Data Collection,17 2020 US Census Demographic and Housing Characteristics file,21 2021-2022 Centers for Medicare & Medicaid Services Provider of Services file,23 and 2020-2021 Maine Rural Health Research Center Ambulance Location Data.15 AD indicates ambulance desert (>25-minute drive from ambulance station); BBD, broadband desert (broadband speeds <100/20 megabits per second); HCD, health care desert (>30-minute drive from health care facility). North Dakota was a noteworthy outlier (Figure 4 and Figure 5). Compared with all 41 states, North Dakota (n = 304 105) had the second-lowest percentage of rural residents living in BBDs (1.4%) and the second-lowest number of rural residents living in BBDs (4368 residents) despite being the state with the tenth-highest percentage of rural residents (39.0%). Intersection of Rural ADs and HCDs With Sufficient Broadband by State Among the 10 states with the greatest number of people living in an AD and HCD intersection with sufficient broadband available, 7 were in the South (North Carolina, Tennessee, Virginia, Kentucky, Arkansas, Texas, and Florida) and 3 were in the West (Montana, Washington, and Oregon) (Figure 4). Among states with the highest percentages of rural populations living in an AD and HCD intersection and sufficient broadband available, 6 were in the West (Nevada, Montana, Utah, New Mexico, Wyoming, and Oregon), 2 were in the Midwest (North and South Dakota), and 2 were in the Northeast (Maine and Vermont) (Figure 5). Intersection of Rural BBDs, ADs, and HCDs by State Among the 10 states with the greatest number of people living in a BBD, AD, and HCD intersection, 5 were in the West (Washington, Montana, Oregon, Idaho, and New Mexico) and 5 were in the South (Virginia, Texas, North Carolina, Alabama, and Florida) (Figure 4). Among the states with the highest percentages of rural residents living in a BBD, AD, and HCD intersection, 8 were in the West (Wyoming, Montana, New Mexico, Idaho, Nevada, Utah, Washington, and Oregon), 1 was in the Midwest (South Dakota), and 1 was in the Northeast (Vermont) (Figure 5). New Hampshire (n = 574 109) was an outlier, with only 2200 rural residents (0.4%) living in an intersection of all 3 deserts, despite ranking among all 41 states as the state with the eighth-highest percentage of rural residents (41.7%). Rural Computing Device Ownership and Broadband Subscription by State Except for 3 Northeastern states (Rhode Island, Connecticut, and Massachusetts), households in rural areas were less likely to have a computing device and broadband subscription compared with urban as well as the overall state mean rate. In 20 states, less than 90.0% of rural households had a computing device and broadband subscription: 12 in the South (all Southern states except Maryland, Virginia, and Delaware), 6 in the Midwest (Michigan, Illinois, Ohio, Indiana, South Dakota, and Missouri), and 2 in the West (Montana and New Mexico) (eTable 1 in Supplement 1). The 17 states with the greatest disparity in broadband subscriptions in rural households with computing devices ranged from percentage differences of 6.0% in Vermont to 12.6% in New Mexico (eTable 1 in the Supplement). Among these 17 states, 2 were in the West (New Mexico and Wyoming), 12 were in the South (all Southern states except Delaware, Maryland, and Virginia), 2 were in the Midwest (Missouri and South Dakota), and 1 was in the Northeast (Vermont). Discussion This cross-sectional study found persistent broadband disparities across 41 states, especially in rural areas of the South and West, where limited internet infrastructure and broadband subscription rates compound challenges in accessing timely medical care.1,3,4,5,6 We estimated that 11.9 million people resided in BBDs, most of whom (88.1%) lived in rural areas; nearly 946 278 rural residents lived in an AD and HCD intersection with sufficient broadband available to support telehealth, and over 649 225 rural residents lived in BBD, AD, and HCD intersections. Additionally, most households owned a computing device, yet rural households had a lower likelihood of having a broadband subscription than urban households. While broadband availability and subscriptions are prerequisites for telehealth, its presence does not inherently guarantee use. These findings suggest that while telehealth has the potential to mitigate gaps in accessing care, its actual impact remains contingent on overcoming a complex set of socioeconomic, health care practitioner and health system offerings, and behavioral barriers beyond mere physical connectivity.1,2,7,9,29 Regional and State-Level Disparities The Western and Southern regions had both the greatest scale and severity of broadband deficits. In the Western region, Washington, Montana, Oregon, Idaho, and New Mexico had not only large rural populations but also some of the highest percentages of rural residents living in BBDs, often exceeding 30% of rural households. The Southern region, home to over half of all residents in BBDs, reflected a similar pattern, with many states facing the dual challenge of underserved broadband and limited local health infrastructure. These patterns suggest that regional strategies must focus on both broadband expansion and strengthening local emergency and clinical care systems. States Leading the Way In contrast, states such as North Dakota and New Hampshire have demonstrated that targeted investments with policy alignment can effectively bridge the broadband gap. Near-universal fiber connectivity in North Dakota, achieved through cooperative broadband initiatives, state leadership, and early federal investments from programs like the American Recovery and Reinvestment Act,30 offers a model for other rural states.31 Similarly, New Hampshire has leveraged federal funds and local partnerships, including electric cooperatives and communication districts, to rapidly expand high-speed internet to underserved communities.32,33,34 These success stories highlight the importance of sustained funding, community-driven networks, and state-level planning that includes assessment of burdensome regulations to enable universal broadband coverage.30,31,32,33,34,35 Telehealth as a Bridge for Care Gaps Telehealth holds promise in regions where ADs and HCDs intersect with sufficient broadband by reducing travel burdens, expanding access to specialists, and improving chronic disease management.10,15,16 Our findings showed that 1.3 million people lived in areas where telehealth could effectively supplement local emergency and clinical services particularly in the South (North Carolina, Tennessee, Virginia, Kentucky, Arkansas, Texas, and Florida) and the West (Montana, Washington, and Oregon). These areas represent a strategic opportunity for state Medicaid, Medicare, and private insurers to support virtual care through policies, such as reimbursement for audio-only visits and expanded health care networks. The telehealth extensions granted by the Consolidated Appropriations Act 2026 especially benefit rural areas by allowing FQHCs and RHCs to serve as Medicare distant site clinics for nontelebehavioral health services delivered via audio-only platforms until December 31, 2027.36 For telebehavioral health services, FQHCs and RHCs can permanently serve as Medicare distant clinics with no geographic restrictions. Furthermore, Medicare beneficiaries can permanently receive telebehavioral health services in their home and use audio-only platforms. Limitations This study has several limitations. First, our estimation of populations living in BBDs assumed that the population was evenly distributed across BSLs; thus, the percentage of underserved and unserved BSLs multiplied by the total population may not have provided an accurate representation. However, assuming a uniform distribution minimized the risk of biasing our results. Second, we used a 50% or greater cutoff to classify a block group as rural, which may not capture mixed urban-rural areas. Given that the census block is the smallest geographic unit available, the measurement error associated with rural-urban classifications at the next level up, the block group, should be minimal.37 Third, the small sample size of the American Community Survey yields large margins of error and is therefore less precise for small populations, such as those in sparsely populated rural areas.38 Additionally, the application of the differential privacy algorithm to the 2020 census data intentionally interjects noise to protect respondent confidentiality, which increases inaccuracies in less populated areas.39 Lastly, this study represents findings among 41 states, with representation across census regions and rural-urban populations. That said, the findings may not represent the 9 states that were excluded due to missing ambulance location data. Conclusions This cross-sectional study showed that 11.9 million residents in 41 US states, predominantly in rural communities, were disconnected from the high-speed internet necessary for telehealth, compounding the challenges of limited access to ambulance and health services. While many households had the devices needed to engage in telehealth, the lack of sufficient broadband remained a critical barrier. These findings suggest that federal and state initiatives, including targeted infrastructure investments and expanded reimbursement policies, are needed to ensure that telehealth can fulfill its potential to reduce disparities in access to timely and high-quality care. Prioritizing regions where BBDs, ADs, and HCDs overlap should be a public health imperative.