Author ORCID Identifier

Year of Publication


Degree Name

Doctor of Philosophy (PhD)

Document Type

Doctoral Dissertation





First Advisor

Dr. Ellen J. Hahn


Lung cancer is a highly preventable form of cancer. Cigarette smoking is the leading cause of lung cancer followed by radon gas exposure and exposure to secondhand smoke. Kentucky leads the nation in both incidence and mortality from lung cancer. Tobacco use in Kentucky continues to be a major public health concern as nearly one-quarter of adults report current tobacco use and just over one-third of Kentucky homes with children lack rules which prohibit smoking in the home. Radon, a colorless, odorless radioactive gas, occurs naturally from the decay of uranium found in rocks and soil and is harmful when it gets trapped indoors. When inhaled, the radioactive particles deposit in the lungs, irradiating cells in the airways. Indoor radon exposure is a cause of lung cancer among smokers and non-smokers; however, a synergistic effect exists between tobacco smoke exposure and radon on the development of lung cancer, putting those who smoke or are exposed to tobacco smoke at a 10-fold greater risk of developing lung cancer than non-smokers and those not exposed to secondhand smoke. Widespread use of tobacco is likely to blame for the high incidence of lung cancer in Kentucky; however, less is known about the contribution that radon exposure has on lung cancer incidence. Testing one’s home for radon is a primary prevention strategy and is necessary to determine exposure risk. Despite public awareness of radon, the proportion of households that have completed home radon testing remains low, ranging from 3-15% of those surveyed. From a public health perspective, it is important to identify variables associated with completion of home radon testing and address disparities in testing in order to create healthy home environments for all. Particularly given the synergistic effect of tobacco and radon, identifying predictors of home radon testing could be useful for prioritizing the allocation of resources and the development of public policy to address environmental risk and improve the health of Kentuckians. The purpose of the dissertation is to: 1) review the literature on methods utilized to quantify population attributable risk (PAR) of residential radon exposure on the development of lung cancer; 2) explore predictors of home radon testing in rural Appalachia; and 3) examine the association between county-level social determinants of health and environmental exposures and rates of home radon testing.

This dissertation has three components; a systematic review of the literature, a prospective study of Appalachian residents, and a secondary analysis of state radon and other population-level data. First, a search of published literature on attributable risk for radon-induced lung cancer was conducted using PubMed for all relevant studies published in English between 2008 and 2018 using the key phrases radon AND attributable risk; lung cancer AND attributable risk; radon AND attributable fraction; lung AND attributable fraction; radon AND population attributable risk. Second, using the Teachable Moment Model (McBride et al., 2003) as a theoretical framework, an exploratory, prospective study design was utilized to examine the association between the Teachable Moment Model constructs and home radon testing in a small sample of rural Appalachia residents. A convenience sample of 58 adult participants was recruited from two rural primary care clinics located in Appalachia Kentucky and were surveyed using a brief paper-and-pencil survey and given a free long-term home radon test kit. Binary logistic regression was used to examine the association between personal risk perception, emotional response, synergistic risk perception and home radon testing. Third, an ecological, descriptive study design was used to conduct a secondary data analysis of 54,683 observed radon values from Kentucky homes. Data from 1995-2016 were obtained from a statewide radon database. Multivariate linear regression was used to examine the association between county-level social determinants of health (e.g., median household income, median home value, percent living below poverty level, percent of the population over the age of 25 with at least a high school diploma, percent owner-occupied housing, and rural-urban status) and environmental exposures (e.g., radon exposure risk potential, adult smoking prevalence, and lung cancer incidence rates) and rates of home radon testing.

Results from the review of the literature revealed four models of excess relative risk typically used to estimate population attributable risk (PAR) associated with indoor radon exposure are described, including those proposed by the Environmental Protection Agency (EPA), the BEIR-VI exposure-age-concentration (EAC) model, the BEIR-VI exposure-age-duration (EAD) model, and the European Pooling Study model. Equations used to calculate PAR vary. Application of the EAC model resulted in higher estimates of PAR. PAR percentages ranged from 5-28% and total number of lung cancer deaths attributable to residential radon exposure ranged from 231-3,366 annually, with more radon attributable lung cancer deaths occurring among those with a history of smoking. When researchers hypothetically reduced radon exposure concentrations in the population, a reduction in radon-induced lung cancer mortality was noted. Uncertainties in estimations stem from differences in approximation of indoor radon concentrations and from use of the two BEIR-VI models since they extrapolate results from the studies of miners to assess lung cancer risk in the general population. Second, results from the prospective study showed that 28 of the 58 (48%) home radon test kits distributed in rural Appalachia were returned for analysis. Eight (29%) exceeded the EPA action level, three of which reported the presence of smoking in the home. Older adults were more likely to complete home radon testing (M= 51, SD= 11 years versus M= 41, SD= 16 years, respectively; p = .008). There were no differences in personal risk perception of lung cancer, lung cancer worry, or synergistic risk perception between those who completed home radon testing and those who did not. Many participants reported low perceived personal risk for lung cancer at baseline despite the fact that 29% of those who tested had high home radon levels. The multiple logistic regression model to determine demographic and personal characteristics predictive of testing status was significant overall, with age as the only significant predictor. Age was associated with completion of radon testing. For every 5-year increase in age, participants were 47% more likely to test their homes for radon. Third, results from the secondary analysis revealed that the average county-level aggregate annual residential testing rate in Kentucky was 13.4 per 10,000 households. Multiple linear regression model to assess predictors of county-level residential radon testing rates was significant overall, with county-level median home value, rural-urban status, upper quartile of the distribution of radon values, and adult smoking prevalence making statistically significant unique contributions to the prediction of residential radon testing rates. For each $10,000 increase in median home values, there was a corresponding increase of 1.54 in the annual rate of residential radon testing per 10,000 households. For every 1-unit increase in RUC value (i.e., an increase in county-level rurality), the rate of annual testing per 10,000 households increased by 1.95. For each additional 1 pCi/L of radon exposure risk potential at the county level, annual rates of residential radon testing increased by 1.36 per 10,000 households. Finally, for each 1% increase in county-level adult smoking prevalence, annual rates of residential radon testing per 10,000 households decreased by 0.50.

Studies to determine attributable risk of radon-induced lung cancer vary in methodology. Given the uncertainties associated with extrapolation of results from miners to the general public, as per the BEIR-VI EAC and EAD models, additional well designed case-control studies using residential radon measurements are needed to provide further evidence for the use of residential radon models over models developed from studies of miners. Providing free home radon test kits as a cue to action in the primary care setting shows promise in prompting radon testing in Appalachia. As radon-induced lung cancer risk increases with exposure over time, efforts are warranted to encourage home radon testing among younger individuals. Additionally, counties with low median home values and high prevalence of adult smoking may benefit the most from public health interventions to increase home radon testing. Public health strategies focused on reducing tobacco and radon exposure in the home are needed in order to reduce the burden of lung cancer and create healthy home environments for all.

Digital Object Identifier (DOI)