The Herald-Leader drew attention to the pressing issue of the undeniably poor health of many people in rural Appalachia, but in ascribing this problem to surface coal mining in the region, the editorial did a disservice to its readers.
As health scientists, my colleague Vanessa Perez and I were asked by the National Mining Association to examine the health effects study by University of West Virginia professor Michael Hendryx that prompted the editorial. We find the nature of the study provides little scientific support for the editorial's strong conclusion.
Hendryx suggested that poor health reported by people living in relatively isolated Appalachian communities correlates to surface coal mining in the vicinity. He did not show that this poor health is caused by surface mining. In fact, the research design used by Hendryx cannot support any conclusions about surface mining's contribution to the incidence of cancer, heart disease, respiratory ailments or other serious health effects.
How we gather scientific evidence often influences the evidence we collect and the uses we can make of it. In this case, Hendryx's team questioned residents in Floyd County coal communities and found that complaints of family health issues were more common there than from people outside of the county.
He did not use health records or pathology reports to verify the recollections of individuals who were asked to recall their own health outcomes or the health outcomes of relatives. This invites misleading conclusions. Someone, for example, could recall that a family member died of cancer who may have resided elsewhere at the time of death. No effort was made to validate or confirm participant responses, which compromises the credibility of the results.
Nor was there any attempt to trace the reported health conditions to mining activity. Just because you live near a power plant doesn't mean you are exposed to its emissions, let alone that the plant's emissions cause your headaches. So even if we can make the assumption that descriptions or recollections about the health of family members were reasonably accurate, there is no evidence mining caused them.
The questioners were not medical professionals, but college students given several hours of training. Given also that their topic is susceptible to emotional reaction, it is probable that bias may easily have crept into the questions, the way they were asked and the responses to them.
A further distortion is likely from the decision to collect information on health symptoms from a single point in time. Analytical epidemiological studies typically require observation over an extended period of time to establish a true pattern from which reliable assumptions can be made. Indeed, an exposure must come before an outcome, but this doesn't mean that an exposure caused or is even associated with the outcome.
Finally, the students questioned family members in other, non-mining communities, with dissimilar incomes, education levels, family history of disease, ages and lifestyles. Advertisers study these demographic factors closely, knowing how they are likely to color our perceptions and habits. People who exercise regularly, watch their diets and avoid smoking may be more likely to participate in surveys and ultimately skew a comparison of responses between communities. Because such individuals are likely to have better health outcomes, comparing responses across the board would be an apples-to-oranges comparison.
None of this is to deny any value of Hendryx's research. His observations can lead reasonable people to form hypotheses that merit valid analytical study. Moreover, he and his colleagues have themselves acknowledged their study's limitations.