About
During the Summer of 2022, DataLab fellows created this dashboard for the South Cumberland Health Network to provide them with the data analysis needed to fill the large primary care health service gaps that are prevelant on the South Cumberland Plateau.
What is DataLab?
Sewanee Datalab is a data science for social good program hosted at Sewanee: The University of The South. This program trains aspiring data scientists that work exclusively on social impact projects partnered with clients and organizations.
Our Community Partner
The South Cumberland Health Network (SCHN) is a non-profit organization that works to remediate barriers to health care access among residents of Grundy county and parts of Franklin and Marion counties on the South Cumberland Plateau of Tennessee. The SCHN serves medically underserved, low-income, and minority populations.
The Fellows
Ellie Davis
Senior at Sewanee: The University of the South; Class of 2023. Majoring in Politics & Women’s & Gender studies.
LinkedIn
Jenna Lusk
Sophomore at Purdue University; Class of 2025. Majoring in Computer Information Technology and minoring in Design & Innovation and Organizational Leadership.
LinkedIn
Kenedi Clinton
Junior at Sewanee: The University of The South; Class of 2024. Majoring in Biology, minoring in Rhetoric, and receiving a Civic and Global Leadership certificate.
LinkedInBackground
Emergency Room Overuse in the United States
Concern has grown in recent years regarding the overuse of emergency departments in the U.S., especially in rural parts of the country where severe ER overuse has been linked to inadequate primary care health service. Investigating trends in ER use can provide helpful insight into the status of the primary care health services of different communities by highlighting who is being underserved and how.
What is Emergency Room Overuse?
ER overuse can be characterized as using the ER to treat non-emergent conditions or ambulatory care sensitive (ACS) conditions. By investigating how frequently patients are going to the ER with a primary diagnosis of an ACS or non-emergent condition, we can quantify ER overuse. ACS and non-emergent conditions can be identified using ICD-10 codes.
The Problem
All three counties that make up the South Cumberland Plateau as defined by the SCHN, (Grundy, Franklin, and Marion) are considered medically underserved by the Health Resources and Services Administration (HRSA). Among other things, this means that the HRSA has designated these counties as having too few primary care providers and services to adequately meet community needs.
Our Project
Our focus this summer is on investigating trends in ER use from the South Cumberland Plateau so as to better understand how the region is medically underserved. More specifically, our analysis provides insight into the demographics of who is primarily “overusing” the ER, and for what specific non-emergent and ACS conditions they are using the ER. By understanding the “who?” and “for what?” of ER overuse, we can better understand the “why?” — why are ERs on the South Cumberland Plateau being overused? What gaps are there between health care services and the needs of the community in this region?
Vocabulary
Non-Emergent/low acuity: Conditions that are non-emergent or low acuity are not emergencies.
Ambulatory Care Sensitive (ACS) Condition: Conditions for which hospitalization can be prevented with consistent outpatient primary care.
ICD-10 Codes: Used to identify conditions and diagnoses.
Primary Diagnosis: Our dataset includes 18 different “Diagnosis” columns for each patient. We have focused our attention on the first diagnosis column for each patient, the primary diagnosis.
Medically Underserved Area (MUA) or Population (MUP): MUA/MUPs are defined by the Health Resources and Services Administration as having a shortage of primary care health services either within an entire geographic area (MUA) or for a population subset within a geographic area (MUP).
ER Overuse by Zip Code
Exploring where ER overuse is coming from provides a great insight into what communities are facing medical gaps the most. The following maps use a scale of blue to red in order to show severity of overuse per county. The maps use blue markers to show the top 10 most visited hospitals by residents of the plateau. The red dots are urgent cares, and the green dots are primary care doctors.
The map visualizes ER overuse on the Plateau broken down by zip code. We found overuse is most prevalent in Sequatchie, TN, with 43.7% of ER visits being instances of overuse. This area is noted as having no urgent care (red dots) nor primary care doctors (green dots). The hospitals shown are the top 10 most visited by patients from the SCP.
Legend
Instructions
Select a zip code for which you'd like to see the top three most visited hospitals from patients from the selected zip code.
Zip Code Key
37301 - Altamont
37305 - Beersheba Springs
37313 - Coalmont
37339 - Gruetli-Laager
37356 - Monteagle
37365 - Palmer
37366 - Pelham
37374 - Sequatchie
37375 - Sewanee
37387 - Tracy City
37397 - Whitwell
ER Overuse by Demographics
Analyzing patient demographics and ER overuse can reveal trends regarding what groups of people face medical gaps or if there is a specific demographic that is significantly underserved. The following graphs show overuse by ACSC and non-emergent by the selected demographics.
Instructions
Begin by selecting a sex, race, and age range*. Then below the black line, you may generate graphs for county, zip code, and insurance type by selecting an option from the drop down menus. You may select multiple zip codes to show on the graph.
* = Ages 70-99 are grouped into one due to Federal Law
ER Overuse by Types of Conditions
Looking into what kinds of ER overuse conditions can help pinpoint what kind of healthcares are lacking in a community. The map and graphs below show the percentage of ER visits that are for ACSC and non-emergent conditions, and also analyze dental, mental health, and substance use conditions to find lack of services for these specific healthcare needs.
The graph indicates overuse for several different kinds of conditions. These conditions can all be treated in other healthcare facilities. This raises questions about a possible deficiency in services within the SCP communities.
Instructions
The map below visualizes the percentage of ER visits that were under certain conditions from each county in the SCP. Select what kind of condition is wanted for analysis and the map will generate. Be mindful that the graph may take a few seconds to load.
Instructions
The graphs below depict the percentage of ER visits were for different kinds of overuse conditions. Select by county, zip code, or insurance types, and the corresponding graph will generate.
ER Overuse by Diagnoses
Analyzing what specific ICD-10 codes are most common amongst the SCP and specific demographics give an exact explanation about what overuse conditions residents are using the ER for.
This graph displays the top 10 diagnoses found in ER visits from SCP residents. These diagnoses are all conditions that could have been treated with primary care or urgent care. This provides insight into the needed health care accessibility for the community to decrease these instances of ER overuse.
Instructions
The following graphs depicts the top 5 ICD-10 codes for the selected sex, insurance type, and county/zip code. To begin, select what sex and what insurance type. Select if you would like to view the graph by county or zip code. Then, select which county/zip code you would like the graph to analyze.
NOTE: Click the following link to look up what a specific ICD-10 code means: ICD-10 Code Search Engine
Primary Findings
ER Overuse
Our data analysis revealed that 40.9% of ER visits from patients in the SCP were instances of ER Overuse, double the national average of 19%.
Overuse by County
Williamson county is the healthiest county in Tennessee. Grundy county, which makes up a large part of the South Cumberland Plateau, is among the least healthy counties in all of Tennessee. This graph compares ER overuse in the South Cumberland Plateau counties with ER overuse in Williamson county. The graph shows that ER overuse is higher in the counties that make up the South Cumberland Plateau than it is in Williamson county.
Overuse by Insurance Type
The data analysis reveals that patients with state/federal government insurance or who self-pay (uninsured) use the ER to treat ACS conditions at a greater rate than those with commercial insurance. The findings may be related to the fact that ERs must treat patients regardless of insurance or lack thereof unlike primary and urgent care clinics.
Overuse by Admit Hour
One of our partner’s initial hypotheses was that there would potentially be a difference in the time of day that patients visited the ER for “appropriate” versus “overuse” reasons. This graph does not support that hypothesis. This lack of evidence supporting the initial hypothesis may indicate that the problem(s) influencing ER overuse on the Plateau lie elsewhere in the system. For instance, is the issue that urgent cares or primary care clinics will not accept certain types of insurance? Are urgent cares in the area too far away, or much farther away than certain ERs? Do the primary care clinics on the Plateau offer all needed health care services?