Author ORCID Identifier

Date Available


Year of Publication


Degree Name

Doctor of Philosophy (PhD)

Document Type

Doctoral Dissertation


Public Health


Epidemiology and Biostatistics

First Advisor

Dr. Wayne Sanderson



Resource extraction exposures are hypothesized to promote adverse respiratory health outcomes among residents of rural Appalachia, yet no studies to date have simultaneously quantified small-scale geographic variation in residential exposure, individual level health factors, and respiratory health outcomes.


The Mountain Air Project (MAP) is a community engaged cross-sectional study based in Harlan and Letcher counties of Southeastern Kentucky. MAP utilized a novel small-scale method to define residential exposure boundaries: hydrologic unit code (HUC), which represents distinct drainages (AKA “hollows”) where residents cluster. We assigned the HUC level density of active and abandoned surface and underground mining, oil/gas wells, coal haul routes, and roadways to each participant. Over a two-year duration, 972 participants with geo-coded addresses and quantified HUC level exposure density were administered a health and exposure survey and a spirometry test by community health workers. In addition, 71 homes received a four-day in-home assessment of gravimetric PM2.5concentration. Particle sampling with direct reading instruments, measuring particle concentrations from 10-10,000 nanometers (#/cc), was conducted among a convenience sample of roadside locations and participant driveways. Multivariable regression models with robust standard errors were used to validate HUC level exposure and to examine the adjusted associations between HUC level exposures, in-home PM2.5, and pulmonary function.


On average, participants were middle-aged (average = 52.9 years of age), female (59%), under-educated (56% ≤ high school education), and above healthy weight (44% obese), with a high prevalence of smoking (33% current smokers; average of 16.1 pack years) and a high prevalence of abnormal pulmonary function (42%). HUC exposure validation models demonstrate that for every one-unit increase in HUC roadway density the adjusted concentration of particles < 100 nm increased by 0.10 log particles/cc (95% CI: 0.03, 0.16) and particles 100-300 nm increased by 0.09 log particles/cc (95% CI: 0.01, 0.17). Participants living in HUCs with the highest tertile of roadway density experienced a 4.3% reduction in adjusted FEV1 percent predicted values (95% CI: -7.44, -1.15) and a 3.8% reduction in adjusted FVC percent predicted values (95% CI: -6.38, -1.21), relative to participants living in the lowest tertile of HUC roadway density. No significant adjusted associations with particle number concentrations or pulmonary function were found among the remaining HUC level environmental exposures. In-home log PM2.5 concentrations were associated with a 24% increase in the adjusted prevalence of abnormal pulmonary function (Prevalence ratio =1.24, 95% CI:1.01-1.51).


Our results suggest roadway and in-home exposures may contribute to adverse respiratory health outcomes among residents of rural Appalachia. Future longitudinal research is needed to further define and quantify HUC level residential environmental exposures and exposure-outcome temporality.

Digital Object Identifier (DOI)