Associations Between Obesity, Physical Inactivity, Healthcare Capacity, and the Built Environment: Geographic Information System Analysis

Figures & data

Figure 1 Study framework.

Table 1 Variables Definition and Data Source

Variables	Definition/Options	Data Source
%Race	African Americans, Whites, Hispanics, and others	Census Population Estimates, 2012
%Gender	Female vs Male	Census Population Estimates, 2012
%Unemployment	Population who are unemployed but seeking work	Bureau of Labor Statistics, 2012
%Uninsured adults	Population under age 65 without health insurance	US Census Bureau’s Small Area Health Insurance Estimates (SAHIE) model 2011, and the American Community Survey (ACS)
%Not seen a doctor due cost	In the past 12 months, needed to see a doctor but could not because of cost.	Behavioral Risk Factor Surveillance System, 2012
% Inadequate Social Support	Adults without social or emotional support	Behavioral Risk Factor Surveillance System, 2012
% Low perception of health	Fair or Poor self-reported health	Behavioral Risk Factor Surveillance System, 2012

Table 2 Demographic and Socioeconomic Profile of South Carolina Counties, 2012

Demographic & Socioeconomic Factors	Overall in SC	SC Counties (n = 46)
		Mean ± SD	Minimum	Maximum
Adult Population (18+ years)
Number (n)	3,643,633	79,209	7866 (Allendale)	355,833 (Greenville)
Percentage of total population			73.8% (Dorchester)	86.5% (McCormick)
Median Household Income	43,290	38,687	25,633 (Allendale)	54,139 (Beaufort)
Female (versus male)	51.4%	50.9(1.8)	45.5% (McCormick)	54.2% (Marion)
Uninsured Adults	24.0%	24.1(2.7)	19.1% (Richland)	33.6% (Jasper)
Unemployed but seeking work	10.2%	10.8(2.8)	6.9% (Lexington)	18.1% (Marion)
Race and Ethnicity
Non-Hispanic African American	27.6%	35.9(16.4)	6.7% (Pickens)	72.9% (Allendale)
Non-Hispanic White	64.0%	57.3(15.2)	23.3% (Allendale)	87.0% (Pickens)
Hispanic	5.2%	4.2(3.0)	1.0% (McCormick, Union)	14.8% (Jasper)
Others	2.0%	3.0(0.9)	0.5% (McCormick, Fairfield)	5.0% (Marlboro)

Table 3 Summary of Study Outcomes and Contextual Factors in South Carolina's Counties, 2012

Characteristics	Overall in SC	SC Counties (n = 46)
		Minimum	Maximum
% Adult Obesity	34.2%	22.5% (Beaufort)	43.9% (Lee)
% Adult Physical Inactivity	28.1%	15.2% (Beaufort)	35.2% (Dillon)
Healthcare facilities (per 100,000)	35.0	4.5 (Marlboro)	116.25 (Charleston)
Primary care physicians (per 100,000)	63.7	11.5 (Williamsburg)	155.0 (Charleston)
Adequate access to exercise opportunities	53.0%	0.05% (Marlboro)	97.3% (McCormick)
Not seen by Doctor due to Cost	16.0%	11.6% (Fairfield)	24.6% (Marlboro)
Inadequate social and emotional support	29.0%	16.7% (Lexington)	39.5% (Allendale)
Low perception of own health	20.0%	10.1% (Beaufort)	28.3% (Marlboro)

Figure 2 Geographic variation in (A) Obesity and (B) Physical Inactivity, in South Carolina's counties, 2012.

Notes: Data from Behavioral Risk Factor Surveillance System (BRFSS).²⁰

Table 4 Results of Correlation Coefficient Analysis

	Healthcare Capacity		Built Environment
Outcomes	Health facilities density	PCP density	Access to Exercise Opportunities
Obesity	(r = 0.408, p = 0.005)	(r = −0.506, p < 0.0001)	(r = −0.603, p < 0.0001)
Physical Inactivity	(r = 0.272, p = 0.067)	(r = −0.392, p = 0.007)	(r = −0.560, p < 0.0001)

Figure 3 Spatial autocorrelation of obesity in South Carolina.

Notes: Data from Behavioral Risk Factor Surveillance System (BRFSS).²⁰

Figure 4 Spatial autocorrelation of physical inactivity in South Carolina.

Notes: Data from Behavioral Risk Factor Surveillance System (BRFSS).²⁰

Table 5 Local Moran’s I Analysis for Spatial Autocorrelation

County	P-value (All Sig.)	z-score	Interpretation
Obesity Spatial Clusters
Greenville (LL)	0.036	2.092	Cluster: has neighbors with similarly low rates (Low rate surrounded by low)
Bamberg (HH)	0.016	2.409	Cluster: has neighbors with similarly high rates (High rate surrounded by high)
Allendale (HH)	0.019	2.336
Hampton (HH)	0.041	2.043
Beaufort (LH)	0.035	−2.099	Outlier: has neighbors with dissimilar data values; (Low rate surrounded by high)
Physical Inactivity Spatial Clusters
Allendale (HH)	0.012	0.012	Cluster: has neighbors with similarly high rates (High rate surrounded by high)
Dillon (HH)	0.008	2.613	Cluster: has neighbors with similarly high rates (High rate surrounded by high)

Figure 5 Spatial autocorrelation of healthcare facilities in South Carolina.

Notes: Data from the SC GIS Data Clearinghouse – DHEC.²⁷

Figure 6 Demonstration of geographic variation in access to exercise opportunities in South Carolina.

Notes: Data from the County Health Rankings.²¹

Figure 7 Demonstration of counties with clusters of low access to exercise opportunities in South Carolina.

Notes: Data from the County Health Rankings.²¹

Table 6 Regression Analysis for Adult Obesity

Characteristics	Unstandardized Coefficients		P-value	Standardized Beta	95% CI
	B	Std. Error			Lower	Upper
(Constant)	20.651	4.251	0.000		12.045	29.257
Physical Inactivity rate	0.448	0.168	0.011	0.395	0.013	0.126
Healthcare facilities (per 100,000)	0.069	0.028	0.017	0.264	−0.061	−0.004
Primary care physicians (per 100,000)	−0.032	0.014	0.025	−0.212	−0.097	0.005
% of adequate access to exercise opportunities	−0.046	0.025	0.075	−0.214	−0.231	0.412
% of not seeing a doctor due to cost	0.090	0.159	0.573	0.068	−0.215	0.367
% of inadequate social and emotional support	0.076	0.144	0.602	0.097	−0.348	0.355
Low perception of own health	0.004	0.173	0.984	0.003	0.109	0.788

Notes: Bold p-values indicates significance at the 0.05 level. Dependent variable: Obesity. Adjusted r²= 0.735, SE = 2.418, F = 18.815, P < 0.0001. Other variables were excluded due to high collinearity (Variance Inflation Factor VIF >4).

Table 7 Regression Analysis for Physical Inactivity

Characteristics	Unstandardized Coefficients		P-value	Standardized Beta	95% CI
	B	Std. Error			Lower	Upper
(Constant)	4.649	4.741	0.333		−4.949	14.246
Obesity rate	0.353	0.132	0.011	0.400	0.086	0.620
Healthcare facilities (per 100,000)	−0.020	0.026	0.446	−0.088	−0.074	0.033
Primary care physicians (per 100,000)	0.004	0.013	0.776	0.028	−0.023	0.030
% of adequate access to exercise opportunities	0.000	0.023	0.986	−0.002	−0.048	0.047
% of not seeing a doctor due to cost	−0.039	0.141	0.783	−0.033	−0.325	0.247
% of inadequate social and emotional support	0.093	0.127	0.468	0.135	−0.164	0.350
Low perception of own health	0.487	0.132	0.001	0.499	0.219	0.754

Notes: Bold p-values indicates significance at the 0.05 level. Dependent variable: Physical Inactivity. Adjusted r²= 0.733, SE= 2.428, F=18.613, P < 0.0001. Other variables were excluded due to high collinearity (Variance Inflation Factor VIF >5).

Centers for Disease Control and Prevention. Behavioral risk factor surveillance system. Available from: https://www.cdc.gov/brfss/index.html. Accessed March 2, 2022.

 Google Scholar

South Carolina Department of Health and Environmental Control. SC GIS data clearinghouse. Available from: https://sc-department-of-health-and-environmental-control-gis-sc-dhec.hub.arcgis.com/. Accessed March 2, 2022.

 Google Scholar

County health rankings & roadmaps. Available from: https://www.countyhealthrankings.org/. Accessed March 2, 2022.

 Google Scholar