Source: UNIVERSITY OF NEVADA submitted to NRP
PREFERENCES OR ACCESS: THE EFFECTS OF BROADBAND INTERNET ON RURAL HEALTH OUTCOMES
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
ACTIVE
Funding Source
Reporting Frequency
Annual
Accession No.
1030658
Grant No.
2023-68006-40056
Cumulative Award Amt.
$649,988.00
Proposal No.
2022-10196
Multistate No.
(N/A)
Project Start Date
Jul 1, 2023
Project End Date
Jun 30, 2027
Grant Year
2023
Program Code
[A1661]- Innovation for Rural Entrepreneurs and Communities
Recipient Organization
UNIVERSITY OF NEVADA
(N/A)
RENO,NV 89557
Performing Department
(N/A)
Non Technical Summary
Stay-at-home policies during the COVID-19 pandemic catalyzed the need for, and the availability of, medical services and educational services that could be delivered online. Thispivot cemented the fact that "broadband access" has become a socio-economic determinant ofhealth. Rural areas in the United States continue to have low broadband access and worse health outcomes in spite of this fact. In this project, we determine how broadband access can be leveraged to alleviate rural health disparities. Our project scope will address goals related to improving policy design, improving broadband access, targeting nutrition security, and assessing the impacts of COVID-19. To achieve these goals, we estimate the causal impact of broadband access on rural health outcomes using a unique panel dataset assembled from publicly available data. We will also determine whether the key constraint in broadband access relates to its lack of demand in rural areas. To do so, we will work with Extension educators to survey rural Nevadans to assess their demand preferences for broadband use. We will also assess the effectiveness of delivering e-services and e-programs. With our interdisciplinary team of economists, nutritionists, social psychologists, agricultural Extension experts, and survey experts, we will develop policy suggestions on drivers of demand for e-services and broadband access. Results from this project promise to provide federal, state, and local stakeholders working with rural communities with insights into the key constraints of broadband use and its impacts on rural health.-at-home policies during the COVID-19 pandemic catalyzed the need for, and the
Animal Health Component
75%
Research Effort Categories
Basic
25%
Applied
75%
Developmental
0%
Classification

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
6076010301080%
7036010301020%
Goals / Objectives
Our project goal is to explore how broadband access can be utilized to alleviate rural health disparities and determine the relative importance of supply and demand constraints for broadband services. We summarize our objectives below and provide a more detailed explanation.Objective 1: Create a unique and comprehensive panel data of health outcomes, socio-economic determinants of health, and broadband access.Objective 2: Assessing the impact of broadband access on rural health outcomesObjective 3: Evaluate the impact of broadband access use in rural Nevada and implement policy workshopsDespite its importance, there are no reliable estimates of the causal effect of broadband access on the health outcomes of rural residents using national-level data. Additionally, even less known about those who still do not have broadband access and how they can be reached or targeted for social programs that can be otherwise delivered online. In this project, we will explore how broadband access can be utilized to alleviate rural health disparities. Our project scope will cover topics related to rural economic development, including policy design, broadband access, and impact, targeting nutrition security, and impacts of COVID-19.Our first objective is to construct a unique and comprehensive panel data of health outcomes, socio-determinants of health (SDOH), and broadband access. We will utilize this dataset to achieve our second objective - assess the effect of broadband access on rural health outcomes. These two objectives would provide federal, state, and local policymakers, as well as academic communities working with rural communities, with insights into how broadband can be leveraged to reduce rural-urban health disparities. In addition, stakeholders like the USDA, state and local policymakers, the FCC, and Extension educators can utilize this data when planning new programs. This dataset would also allow for comparisons over time, a unique opportunity to measure progress.For the third objective, we will work with our interdisciplinary team of economists, nutritionists, social psychologists, Extension experts, and survey experts based at the University of Nevada Reno (UNR) to assess the state of broadband supply and its use in rural Nevada. To separate the relative importance of supply and demand constraints for delivering e-services and e-programs to rural Nevadans, we will first work with Extension educators to understand current grassroots issues in delivering e-services. We will then survey a random sample of rural Nevadans to assess their preferences for broadband use and understand their internet use patterns and the supply constraints they face. The advent of COVID-19 forced many public and private programs online. For example, in Nevada (as in many other states), the SNAP-Ed program pivoted to the online format as COVID-19 restrictions at the state level prevented regular gatherings at the local community centers. This pivot provides us a unique opportunity to test the relative importance of the quality of broadband supply vis-à-vis preference for more in-person programming among SNAP-Ed households in rural Nevada in explaining the demand for e-programs and e-services. This exercise would inform future program designs of e-services and e-programs.Meeting these objectives would support our project goal to explore the sustainability of delivering agriculture and food services through the use of broadband access, especially while considering the implications for reaching areas and individuals in most need.
Project Methods
The Methods for each objective are described below:Objective 1: Create a unique and comprehensive panel of health outcomes, socio-economic determinants of health, and broadband access.We will construct a panel dataset with measures of broadband access, including annual measures of availability and the speed of broadband at the block level. The panel for health outcomes and socio-determinants of health (SDOH) will depend on the geographic granularity of the availability of secondary data. We will merge broadband access data available at the census block level with health outcomes and SDOH data (at the census tract or county level) to create the panel.Objective 2: Assessing the impact of broadband access on rural health outcomesUsing the data from Objective 1, we will estimate the marginal effects of changes in broadband access on the changes in rural health disparities over the sample period using exogenous variation generated by policy changes to broadband delivery. Our goal is to explore whether a causal relationship exists between broadband access and health outcomes.Our starting point is a fixed effects model where we follow one geographic unit (census block, block group, tract, or county level) over time. Regressions will be clustered at the relevant geographical unit. This methodology will allow us to explore the marginal impact of broadband access on health outcomes.Unobserved heterogeneity may make broadband access variable endogenous. For example, internet service providers are likely to expand in high-growth and high-income areas, while at the same time, residents from affluent areas are more likely to have better health outcomes. To address this, we will include extensive SDOH variables, from the dataset constructed above to account for cross-county differences (such as the percentage of the population in different age groups, race, ethnicity, education, income, housing, etc.).Policy changes cataloged by the National Conference of State Legislatures (NCSL) generate plausibly exogenous changes in broadband access that may provide a consistent estimate of the effect of broadband access on health outcomes. For example, some states could have made it easier to adopt broadband technologies by subsidizing use, while others may have encouraged the growth of mobile phone coverage instead. Alternatively, rural infrastructure development projects provided incentives to upgrade last-mile connectivity using fiber optic cables at different times in different states. Hence some areas might have had access to fast internet services quicker than others. Differences in policy timing can be utilized to generate causal estimates using a DID estimation such as the Two-Way Fixed Effects (TWFE) model.In our project we will compare the availability of internet services providers between states where municipalities were allowed to provide a public option with those that did not allow any government participation (such as Nevada). Conditional on changes in removing variation that is common within a state and over time, we will utilize exogenous variation of policy implementation to causally estimate the impacts of broadband use on health. For this, we will utilize a) a reduced form two-way fixed effects (TWFE) model, and b) an instrumental variable (IV) estimation. Recent advances in this literature have shown that TWFE model estimates maybe biased (Goodman-Bacon 2021). Hence, after identifying the policy of interest, we will use updated methods by Calloway and Santana (2020).We will use a similar strategy to compare if the type of technology (mobile broadband access vs. fixed-line access) or the quality of internet (high-speed or low-speed services) matter more to health. By using within-county changes in broadband access to estimate changes in rural health outcomes over time, we will estimate a dollar value of the marginal impacts. In all cases, we will compare counties with and without persistent poverty, counties with above median minority populations, and counties with high or low shares of elderly adults, to derive implications for the role of SDOH. Since broadband use may vary by age, with younger individuals using services at much higher rates than older individuals, we will compare if more internet-time-exposure[1] leads to better outcomes as compared to those who acquire broadband access at a later age within a county.Objective 3: Evaluate the factors behind broadband adoption in rural Nevada and implement policy workshops The SNAP-Ed program is conducted by Extension. Nevada's Extension educators (currently 14 in total) are its key trainers and facilitators. They have deep local knowledge of the challenges and opportunities for implementing similar programs. Many educators have firsthand experience in pivoting the SNAP-ed program to the online format when COVID-19 restricted access to their respective communities. Hence the first task in meeting objective 3 would be to conduct a Zoom workshop (in the second year) to leverage the expertise and knowledge of these educators. Results from Objective 2 will be presented to inform educators of the current research project and the direction of the proposed survey.After insights have been gathered, the integrated research and extension team will collaborate with the Nevada Center for Surveys, Evaluation, and Statistics (NCSES) to survey a random sample of 600 rural Nevada residents. Leveraging the expertise of NCSES and the knowledge collected from Extension educators, a 10 to 20-minute bilingual survey (English and Spanish) will be developed to inquire about how individuals prefer to access the internet (fixed-line or smartphone access), perceptions and satisfaction of the internet speed, preferences for use of internet-related services (e.g., applications for emails, video chats, social media use and comfort with ordering goods online), basic health outcomes, and key demographics. For those who attended the SNAP-Ed online program, a short assessment of the quality of SNAP-Ed's e-programs will also be conducted, including their perceptions of the online format and if they completed the program. Eligible individuals who did not participate will be asked why they did not participate. The NCSES team will disseminate a multi-modal survey developed in task 3.2 in year 3. NCSES will purchase an address-based list sample from Marketing Systems Group to develop a random sample of residents from the 14 rural counties in Nevada. Once data is collected and analyzed, In addition to this multi-mode design using both web and paper, a design we have used for many surveys at NCSES that has higher response rates than a paper-only survey (McMaster et al. 2017), we will also use a surveillance approach. We will utilize maximum likelihood estimation models to explore the roles of broadband availability and preferences in facilitating program participation in online training and education programs.[1] Internet time exposure is the number of years that someone had access to broadband in their county, number of years an individual had access to high-speed internet, or the number of years that the person had additional technology at the county level.

Progress 07/01/23 to 06/30/24

Outputs
Target Audience:In the first year, researchers spent time gathering and collating data on rural broadband access and rural health across the United States. The key audiences are academics and policymakers for whom this data might be useful. Once the dataset is complete, we plan to make this dataset publicly available. We anticipate the dataset to be useful to individuals working on the topic of broadband including industry professionals, government researchers, academic faculty, and students. Extension educators who could learn from this research will be approached over the course of the project. We have received feedback from the advisory committee on the ongoing work and preliminary results and we will be working on submitting publications from this project for the next year. Changes/Problems:Co-PI, Dr. Katherine Lacy, has moved to the USDA Economic Research Service in July 2023. Co-PI, Dr. Anaka Aiyar, moved to the University of Vermont in August 2023. She is now the PI of a sub-award that was transferred to the University of Vermont in May 2024 Even though their affiliations have changed their roles in this project have not. Both of them are working and will continue to work as described in the project initiation report. What opportunities for training and professional development has the project provided?Trainingactivities included advising graduate students on how to: gather, clean, and compile datasets. conduct a literature survey. This activity was conducted both in Nevada as well as in Vermont. conduct data analysis. write draft reports. Professional development activities - Attending the Agricultural & Applied Economics Association annual conference to network and collaborate with other Principal Investigators. How have the results been disseminated to communities of interest? Nothing Reported What do you plan to do during the next reporting period to accomplish the goals?During the next reporting period (July 2024 to June 2025) we will primarily focus on the second objective. Objective 2: Assessing the impact of broadband access on rural health outcomes Task 2.1: Panel estimation to measure how changes in broadband access impact rural health outcomes Task 2.2: Establishing causality in the relationship conditional on time-variant controls Task 2.3: Establishing causality using a two-way fixed effects model Task 2.4: Establishing causality using an instrumental variable Task 2.5: Exploring heterogeneity in data to derive policy implications We plan to present at conferences andsend out papers for possible publication in peer-reviewed journals. We will also lay the groundwork for the third objective. We will conductfocus group discussions, complete the IRB process, and develop the questionnaire for the survey.

Impacts
What was accomplished under these goals? We completed the first objective of the project (Create acomprehensive panel data of health outcomes, socio-economic determinants of health, and broadband access). We are currently working on the second objective. We held regular team meetings to discuss the data-gathering and cleaning process. These meetings happened about once a month. A stakeholder advisory board (SAB) meeting was conducted in June 2024. Objective 1: Activities: We have gathered dataon broadband speed, loans and grants given to rural areas to expand broadband,and health outcomes from 2017-2021. Data gathered will be launched on a website for the general public,policymakers, government officials, and academic communities. Details of the website will be available soon. We use census tract-level data from 2017 to 2021. The number of census-tract-year observations is 355,521. The data on Broadband speed show that the average download speed during our sample period was 160.6 MBPS and the average upload speed was 54.5 MBPS. The median download and upload speeds were lower (23.6 and 2.6 MBPS respectively) and the average number of providers per Census tract was 11. As for health outcomes, we examine several outcomes such as Obesity, Stroke, COPD, Asthma, Cancer, Depression, and Smoking. Key Outcomes:We have completed merging multiple national-level datasets of health outcomes, socio-economic determinants of health, and broadband access based on geographic granularity (such as county or census tract). The data set will be updated every year until the end of the project. Audience targeted for impact - researchers, policymakers, Extension educators, and graduate students.

Publications