eHealthEquity

Active Research

Research Notebooks, Data Sources, and outputs of eHealthEquity research.

Dissertation Research of Matthew Farmer MSN, RN University of Missouri, Sinclair School of Nursing @mfarme

Aim

Assess and analyze quantifiable evidence that Internet access and adoption is a social determinant of health in United States counties and describe the interaction effects between other social determinants.

• Why publish on Github? To promote replication with different or new data, grow research in this domain, and do science in the open. Reference: Coducting Replication and Repoduction in Science

Research Question 1

• What is the relationship between Internet adoption and self-reported diabetes prevalence in United States counties, controlling for other social determinants of health?

Research Question 2

• Do machine learning algorithms produce a better fit (lower variance and bias) compared to linear regression models?

---

Libraries Used

• Pandas
• Numpy
• cenususdis
• sodapy
• scipy
• matplotlib
• seaborn
• geopandas
• shapely
• SciPy
• scikit-learn
• statsmodels
• plotly
• os
• json
• urllib

Data Sources and Variables

Variables	Level	Concept	Measure	Source	Note
RQ1 DV	Outcome	Diabetes Prevalence	Diagnosed diabetes among adults aged >=18 years; % of county	BRFSS/PLACES 2021	Upper and Lower confidence; crude and age-adjusted
IV	Intermediary	Behaviors and Biological Factors: Internet Adoption	Households with internet (Broadband) subscription; %	ACS 5Y 2021	proportion and raw count with margin of error
Covariates	Structural	Gender, Age, Ethnicity	Aged 65 & Older, Gender, Race, Ethnicity	ACS 5Y 2021	proportion and raw count with margin of error
Covariates	Structural	Education, Occupation, Income	Educational attainment, Industry employed, employment status, income in past 12 months below poverty level	ACS 5Y 2021	proportion and raw count with margin of error
Covariates	Intermediary	Material Circumstances	Transportation, Housing type, Health insurance coverage	ACS 5Y 2021; BRFSS 2021 (insurance)	proportion and raw count with margin of error
Covariates	Intermediary	Behaviors and Biological Factors	Obesity, smoking status, visit to doctor in past year (adults)	BRFSS/PLACES 2021	Upper and Lower confidence; crude and age-adjusted
Covariates	Intermediary	Psychological Factors	Depression among adults, Mental health “not good” for >= 14 days among adults	BRFSS/PLACES 2021	Upper and Lower confidence; crude and age-adjusted

---

Data Sources

• American Community Survey (ACS)
  • ACS 5-Year Data Profiles
  • ACS 5-Year Data Profiles API
• BRFSS/PLACES
  • BRFSS/PLACES API
• NTIA
• RUCA Codes

Conceptual Framework

• WHO Social Determinants of Health Framework