Again, lower socioeconomic status is believed to shape healthcare experiences and health outcomes. In this post I want to understand how a particular R package calculates ADIs, and then also see a few alternative approaches before exploring how things look for Ohio in particular and then perhaps for Appalachia, more generally.
“SDOH, although experienced by individuals, exist at the community level. Healthcare systems that learn about the communities their patients live in, and the community-level barriers members can face to becoming and staying healthy, can better adapt their recommendations to people’s lives.” In line with the necessity to learn about social challenges and then develop solutions or workarounds that ultimately ought to improve the health of communities, the Agency for Healthcare Research and Quality (AHRQ) has a social determinants of health-related beta running whereby they provide annual data series (through 2018)1 for counties and zip code tabulation areas (ZCTAs).2
AHRQ puts SDOH into five key areas:
| Area | Description |
|---|---|
| Social context | (e.g., demographics, social networks and supports; social cohesion; racial, ethnic, religious, and gender discrimination; community safety; criminal justice climate; civil participation). |
| Economic context | (e.g., employment, income, poverty). |
| Education | (e.g., quality of day care, schools, and adult education; literacy and high school graduation rates; English proficiency). |
| Physical infrastructure | (e.g., housing, transportation, workplace safety, food availability, parks and other recreational facilities, environmental conditions, sufficiency of social services). |
| Healthcare context | (e.g., access to high-quality, culturally and linguistically appropriate, and health literate care; access to insurance; healthcare laws; health promotion initiatives; supply side of services; attitudes towards healthcare; and use of services). |
The AHRQ also emphasizes social needs, arguing that “While everyone who lives in a community shares exposure to the same SDOH, individuals have varying social needs. For example, one member of the community might be homeless, while another has adequate housing. Increasingly, healthcare systems are trying to assess the specific social needs of their patients and help meet those needs.” What needs might these be? The following list is offered up by the AHRQ:
| Social Need | Description |
|---|---|
| Social support | (e.g., social isolation). |
| Communication barriers | (e.g., hearing or vision impairment, lack of English proficiency). |
| Trauma | (e.g., adverse childhood experiences, domestic violence, elder abuse). |
| Educational barriers | (e.g., learning difficulties, limited literacy). |
| Food insecurity | (e.g., going hungry, worrying that you won’t have enough food). |
| Housing insecurity | (e.g., homelessness; living in overcrowded, unsafe, or unstable conditions). |
| Financial strain | (e.g., being unable to pay for medicine and other essentials). |
| Employment insecurity | (e.g., being un- or under-employed). |
| Lack of access to legal services | (e.g., combat discrimination, unsafe workplace or housing, criminal defense, immigration status, victim or protection services, guardianship or custody). |
| Lack of transportation | (e.g., inability to get to workplace or healthcare sites). |
| Physical environment | (e.g., lead paint). |
The data files contain variables that correspond to the five key SDOH domains highlighted earlier: social context, economic context, education, physical infrastructure, and healthcare context. And given the geographies,additional data can be joined to these data for counties and ZCTAs. Data documentation is available here as a pdf file and a variable codebook (also shown below) is available here as an xlsx file
The County and the ZCTA files have a total of 238 and 166 indicators, respectively. Details of the variables can be seen below.
County Indicators (2018)
ZCTA Indicators (2018)
I will pick a couple of indicators to explore, and please remember that these choices are driven simply by the rarity of seeing these measures displayed in the wild. The few I will focus on are all about food access, transportation, access to healthcare services, and some features of the healthcare system in the geography.
Since the data are drawn from multiple sources, suppression rules (if applicable at source) will have to be borne in mind. For example, anything coming from the County Business Pattern data series will have no data if the number of establishments number three or less. In our state of Ohio, for example, many of the counties have community food service organizations but these are too few in number to avoid suppression, leading to low utility of this particular indicator for most of our counties.
Well, turns out there are few measures from the County Business Pattern that will cover all counties. As such, food access may be a challenged set of indicators, at least at the county-level. If this is the situation with 3,224 counties, then it makes no sense to expect there to be no suppression at the ZCTA level, given the 33,120 ZCTAs. Looking at dental, mental, and primary care access might be an easier task since these are ordinal measures and hence available for every county. How do Ohio’s counties fare? Well, here we are …
Shortage of Dentists
Shortage of Primary Care Physicians
Shortage of Mental Healthcare Providers
It is evident that most of Ohio’s committees face shortages of dentists and primary healthcare physicians, but mental healthcare providers are less scarce. This latter observation is a welcome reality during the COVID-19 pandemic. Southeast Ohio continues to pay an unusual access penalty, especially for dentists. This forces many Ohioans, particularly those without employer-sponsored health insurance, to cross county borders to get dental care for themselves or their children. This is a ridiculous state of affairs in a so called “world superpower.”
AHRQ conducted an environmental scan of SDOH data sources to inform the development of AHRQ’s SDOH beta data files publicly released in December 2020. AHRQ is releasing the environmental scan as a resource for analysts interested in identifying SDOH-related data sources. The scan aims to identify as many SDOH data sources as possible as of July 2020 at the ZIP Code, county, and State level for the domains of social context, economic context, education, physical infrastructure, and healthcare context. The scan is organized in an Excel spreadsheet to maintain the filter functionality of each column in the scan so that analysts can sort by variables such as SDOH domain or level of geography. This environmental scan is available here as an xlsx file↩︎
These files contain the beta version of AHRQ’s database on Social Determinants of Health (SDOH), which was created under a project funded by the Patient Centered Outcomes Research (PCOR) Trust Fund. These SDOH beta data files are curated from existing Federal datasets and other publicly available data sources. The purpose of these files is to make it easier to find a range of well documented, readily linkable SDOH variables across domains without having to access multiple source files, facilitating SDOH research and analysis. The SDOH beta data files are a starting point to inform additional work AHRQ is undertaking to develop SDOH data. Source↩︎