Recipient Organization
UNIVERSITY OF WYOMING
1000 E UNIVERSITY AVE DEPARTMENT 3434
LARAMIE,WY 82071-2000
Performing Department
(N/A)
Non Technical Summary
This integrated New Investigator project addresses the "agricultural economics and rural communities - rural economic development" priority area by investigating how the consolidating healthcare industry has influenced rural health and economic outcomes. We leverage experienced perspectives from rural healthcare stakeholders to inform and develop a detailed analysis of restricted-access government databases. Project outputs include Extension reports, online tools, and legislative policy briefs.In addition to its contributions to individual well-being, healthcare is a major economic driver, crucial to the vitality of rural communities and industries. Unfortunately, publicly available datasets for rural areas are heavily suppressed for data privacy leading to the extant literature on this topic being urban focused or limited to a specific geography. Meanwhile, hundreds of rural hospitals have closed while rural health outcomes continue to lag.The project begins by engaging its diverse and experienced advisory board in industry and community trends, target audiences, evaluation methods. The project team will then conduct online roundtables to hear from community stakeholders about current barriers and opportunities facing their local healthcare industry. The roundtables will also explore how changes in healthcare services have influenced local health outcomes. Subsequently, researchers will access restricted establishment and health data via the Federal Statistical Research Data Centerto explore stakeholder insights, assess the human costs of healthcare's consolidation, and attempt to identify imperfect substitutes for critical healthcare services. The project will conclude with the development and disemination of Extension materials to target audeinces.
Animal Health Component
80%
Research Effort Categories
Basic
20%
Applied
80%
Developmental
(N/A)
Goals / Objectives
The long-term goal is to narrow the gap between rural and urban residents' health outcomes by ameliorating access to care issues that are increasingly common in rural communities. This will be achieved by accomplishing the following objectives:1. Predict the incidence of specific health outcomes and possible trends at the individual and county level based on the prevalence of multiple types of healthcare establishments (physicians, women's health clinics, hospitals, etc.), prevalence and quality of employer contributions to health insurance policies, and the share of Medicaid/Medicare recipients.2. Examine the influence of community assets and industries on rural health outcomes.3. Evaluate how structural changes in the healthcare sector (e.g., hospital closures and consolidations) have impacted rural residents' health outcomes and access to care.4. Create and implement an Extension "at-risk" community program to identify potential barriers and threats to healthcare accessibility, piloting in partnering state counties (Wyoming, Utah, and Michigan), and then, after stakeholder involvement and associated revisions, extending the program nationwide using an online train-the-trainer model.
Project Methods
Research:The research team will first merge county aggregated Vital Statistics data on patient outcomes with the Longitudinal Business Database (LBD), Integrated Longitudinal Business Database (ILBD), and the American Community Survey (ACS), to produce a panel of restricted- access datasets spanning the years 1990-present across all counties in the contiguous US. This will then be merged with publicly available data from the Behavioral Risk Factor Surveillance System, Centers for Medicare & Medicaid Services (CMS) Open Payments Data, Food and Drug Administration National Drug Code Database, and the CMS Medicare Provider Utilization and Payment Data: Part D Prescriber, amongst others. The resulting dataset will contain a wealth of information on county health outcomes, locations and financial health of healthcare establishments, residents' usage of different insurance programs, telehealth usage or availability, as well as environmental and behavioral risk factors. Two important pieces of this data that are a result of their restricted-access nature are 1) the establishmentfirm identifiers and 2) the patients' place of residence and place of care. The first enables researchers to analyze firm sizes, efficiency changes from firm mergers or acquisitions, and overall local market concentration. The second enables researchers to observe how far patients traveled to receive care for their health outcomes. The LBD, ILBD, and CMS datasets provide detailed information on sources of physical and financial access to healthcare and will greatly contribute to the literature. Specifically, the datasets allow for an analysis that includes a comprehensive geography (the contiguous 48 states), multiple sources of healthcare within sectors and communities, and a large enough time frame to appropriately identify the causal structure of healthcare access on community health and economic development outcomes. Given the geographical scope of these datasets, the researchers will also be able to explore how healthcare access and outcomes differ across the rural-urban spectrum (as defined by the USDA Rural-Urban Continuum codes). It is important to note that our research hypotheses and statistical models will be informed by the advisory board.The primary method of analysis will be panel data models. Various distributions will be tested to find an appropriate distribution for each model depending on the dependent variable and specific research question. Given the nature of establishment data, count data models such as the Poisson, Negative Binomial, and their zero-inflated counterparts are expected to be most appropriate. When the dependent variable is a continuous measure such as the percent of residents with a certain health outcome a simple normal distribution for county-level analyses may be most appropriate. Fixed effects will be used where appropriate to control for time invariant factors.When feasible, researchers will apply spatial weights matrices to account for spatial relationships. For example, when the dependent variable is the incidence rate of a specific health outcome, researchers will use the Spatial Durbin Model (SDM), Spatial Autocorrelation Regression (SAR), Spatial Lag of X (SLX), and/or the Spatial Error Model (SEM). The proposed analysis must control for potential estimation issues. Some econometric ailments such as heteroskedasticity and multicollinearity may be simply corrected for using statistical tests and corrections; however other issues such as endogeneity must be accounted for by testing coefficient sensitivity using independent variable lags and testing instrumental variables. In some cases, variables may be endogenized within a system of equations. The most probable challenges in estimating the panel data models are expected to be identifying the lagged structure of the healthcare establishments for non-emergent health outcomes (e.g., cancer or cesarean sections) and addressing endogeneity. However, it should be noted that knowing patients' places of residence and care will reduce the endogeneity problem of counties with hospitals observing more deaths.Extension:Extension activities at the beginning of the project include virtual roundtables with industry and community stakeholders to assess needs, barriers, and opportunities. These virtual roundtables will be hosted using password protected Zoom virtual meeting rooms hosted by the University of Wyoming. Stakeholders will be identified and recruited through external partners' networks, including the networks of those serving on the advisory board. The results of the virtual roundtables will then be shared with the advisory board to assess how trends, observations, and concerns may be addressed through the team's research and backend Extension activities. After conferring with the advisory board, the results of these stakeholder discussions will be written up in a report for dissemination to participants, state Extension organizations, and the Cooperative Extension Service. Elements of this report will be used as motivation for suggested actions in future policy briefs. As the research portion of the grant project concludes, grant team members will again meet with the advisory board to share interpreted modeling results and translate them into Extension outputs. The county-level analysis (informed by the other analyses) will be used to create the at- risk mapping tool which will identify counties that are at risk for developing more negative health outcomes based on their current and predicted trajectory of physical, financial, and social access to care. Visually the maps will look like choropleth maps of the contiguous US with darker shades indicating a greater risk to health outcomes based on the county's industries, proximity to healthcare services, and financial access to healthcare. These maps will assist officials, practitioners, and professionals in determining what the county's main threats, barriers, and opportunities to care are in the present and future periods. After an initial pilot period of six months, the tool will be reevaluated by the grant team and advisory board. Reevaluation of the at-risk mapping tool will focus on inquiring who is using the tool, how it is being used, and what information is still needed. These maps will be made by hiring Wyoming Geographic Information Science Center, an interdisciplinary research and education institute with extensive experience in managing and mapping data for educational purposes. After production, PD Van Sandt will manage the maps on the University of Wyoming's website, though the tool will be advertised and accessible through links or embedded interfaces on other websites, as well.In addition to Extension tools and information for stakeholders, the project team will work with the advisory board to develop policy briefs for state and federal legislatures and departments of health. The purpose of these policy briefs will be to summarize and translate our research findings into meaningful information, policy suggestions, and healthcare industry predictions for rural places. The project team will leverage the advisory board's skills in networking and policy development to produce effective policy briefs.Finally, the project team will work with communities to beta test the outreach materials. The team will identify high risk counties in each state to engage in dialog around findings. These meetings will conclude with a list of actionable items that the communities can take action on to address their local healthcare concerns. As these activities expand toward the national scale, we will adopt an online train-the-trainer model to increase capacity and reach.