This is the first person-level study of adult health status on a small geographic scale (i.e., smaller than county) in an environment characterized by mountaintop coal mining. The odds for reporting cancer were twice as high in the mountaintop mining environment compared to the non-mining environment in ways not explained by age, sex, smoking, occupational exposure, or family cancer history. No one type of cancer was responsible for the effect.

Environmental pollution contributes to cancer risk [31, 32, 33, 34, 35], and many chemicals that are present in coal, coal strata, and coal processing activities are established or possible carcinogens. Arsenic, for example, is an impurity present in coal that is implicated in many forms of cancer including that of skin, bladder and kidney [31, 36]. Cadmium is linked to renal cancer [34]. Diesel engines are widely used at mining sites, and diesel fuel is used for surface mining explosives, coal transportation and coal processing; diesel exhaust has been identified as a major environmental contributor to cancer risk [37].

Previous research on Appalachian health disparities has tended to focus on health care access problems, or behavioral risks such as poor diet and smoking, as the causal factors driving poor health outcomes. A recent study in Virginia, for example, identified higher cancer rates among Appalachian compared to non-Appalachian residents, and discussed the need for better health care in Appalachia [5]. However, Appalachian Virginia also has mountaintop coal mining, and the environmental, social and economic impacts of coal mining are often overlooked in Appalachian health research. Mountaintop coal mining is damaging to the environment, and contributes to the area’s chronic economic problems; these areas have the highest poverty rates and highest unemployment rates in the region [9, 38]. Poor economic conditions are one of the most powerful predictors of poor public health outcomes [39, 40].

Study limitations include those relating to survey sampling procedures and the extent of questions asked. Contact attempts at most households occurred only once, and survey times did not include late evenings and weekends because of the logistical and cost difficulties involved in transporting and housing the student volunteers. This could result in survey respondents in both locations that are not necessarily representative of the entire populations. Survey procedures, however, were comparable in both communities and so would not be expected to result in an overestimate of cancer in Coal River relative to Pocahontas. Asking people if they have ever had cancer limits the cancer experience to survivors. Information on persons who died from cancer was not collected, which may explain why some cancer types such as lung cancer were observed rarely in the sample. In other research, higher lung cancer mortality in coal mining portions of Appalachia has been documented [10].

The survey included limited information on covariates. The preference among the community research partners was to keep the survey brief so that as many surveys as possible could be collected in a short time. There were concerns expressed by community partners that if the time spent per survey was prolonged, such that fewer surveys could be completed/day and more time had to be spent in Coal River to collect an adequate sample size, word about the survey taking place would reach the coal industry, and community residents would be instructed or pressured by industry representatives not to take part. Limited covariate data precluded investigating the possible impacts of such variables as obesity or health care access on cancer.

After data collection the research partners convened to discuss how the study process could be improved for possible replication in other communities. There was agreement that future efforts should attend to data collection during weekend and evening times, and how some additional data would have been worth the extra survey time to collect. In addition to extra covariates such as obesity, community partners expressed the importance in future studies of collecting cancer data not just on biologically-related family members but on spouses, as they knew of cases where a husband or wife had recently died of cancer, but these data were not collected in the survey. These discussions illustrate the utility of the community-based participatory model in helping both parties (academic researchers and community residents) learn from each other to make research efforts more practical and effective.

As a partial response to the limited covariate data, we compared Boone, Raleigh and Pocahontas counties on poverty, health care access and obesity rate indicators. Data from the US Department of Agriculture Food Atlas indicate similar adult obesity rates in the three counties: 32.4% in Pocahontas, 31.3% in Raleigh, and 33.6% in Boone [41]. Poverty rates for 2007 as reported in the 2008 Area Resource File were 15.7% for Pocahontas, 16.7% for Raleigh, and 18.2% for Boone [42]. Pocahontas County was not designated as a Health Professions Shortage Area for 2008, Boone County was designated as a shortage area, and Raleigh County was a partial shortage area [40]. Although, these differences indicate that environmental effects are not the only influence on health outcomes, they also highlight the economic and health care problems that are present in disadvantaged mining environments.

Pocahontas County was chosen for the comparison location because it is in southern West Virginia and does not have active coal mining, but otherwise was selected due to personal contacts among the community research partners. Future research that surveys community health in more closely matched mining and non-mining communities could reveal a better understanding of comparative health in coal mining and non-coal mining towns. Further person-level comparisons between communities of traditional coal mining practices versus mountaintop mining practices could also yield better understanding of the effects of coal mining on the health of Appalachian communities.

The higher cancer rates in Coal River cannot be attributed to direct occupational exposures among coal miners. In fact, former or current coal miners who completed the survey did not have higher cancer rates than non-miners, perhaps reflecting a ‘healthy worker’ phenomenon.

The National Cancer Institute (NCI) reported that in 2007 about 3.9% of the US population consisted of cancer survivors (excluding non-melanoma skin cancer) [43]. Rates in the current study excluding non-melanoma skin cancer were 9.9% across groups; that is, there was a high cancer rate in both groups compared to the national average. However, a report published by the state of West Virginia estimated that 10.4% of adults in the state were cancer survivors in 2009 [44], which is close to the rate found in Pocahontas. The state report also says, however, that the US prevalence rate is 9.6% without citing the source for this figure; the national rate in this state document does not agree with the NCI report.