Study Finds Model Used in Activists’ Research Doesn’t Jibe with Real Air Monitoring
A new study raises questions about the model activist researchers use in their attempts to link fracking to health risks.
A series of studies from Johns Hopkins University used a metric the researchers call “well activity” as a proxy for actually measuring exposure to air pollutants they claim are associated with oil and natural gas development. However, nobody had tested whether well activity accurately correlates with real exposure to air pollution.
A new study led by Judy Wendt Hess and a team from Shell, published in the International Journal of Environmental Research and Public Health, tested the assumption that well activity can be used as a reasonable proxy for exposure to air pollution. It found that this foundational assumption is deeply flawed, as the researchers reported:
“The question we essentially asked was, if [Pennsylvania] monitoring sites were instead a sample of epidemiology study subjects’ homes with monitors placed outside the front door, how well does the categorization of exposure agree between the two methods? We found that they did not agree well at all with the same exposure quartile assigned in roughly one in four observations, and the opposite category assigned for roughly 25%.” (emphasis added)
The well activity metric incorporates numerous factors, including the density of wells around a residence, the phase of well development (pad preparation, drilling, hydraulic fracturing, and production), well-depth, and well production volume. Unlike other ‘proximity’ metrics however, it doesn’t limit the study to people with wells near their homes, but rather assumes that every well in the state contributes to the exposure of every resident of the state.
By calculating these metrics as a proxy for exposure at different locations, and then comparing health outcomes for people at these varying levels of exposure, the Johns Hopkins researchers have attempted to connect fracking to asthma, low birth weight, depression and other ailments.
Those studies did not use actual air measurements, though, and until now no study had tested if well activity was a valid method of assigning exposure, or if it could introduce errors of its own. This study was the first to test this basic presumption.
Using data from nearly 80 Pennsylvania Department of Environmental Protection air monitors and the U.S. EPA’s Air Quality System, Hess’s team compiled daily concentrations of fine particulates, nitrogen dioxide, sulfur dioxide, carbon monoxide, ozone and benzene between 2011 and 2015. The monitors were used to mimic residences, and were assigned exposure categories ranging from ‘very low’ to ‘high’ every day of the study period based on the air data. They then established which unconventional wells were operational during that period and assigned exposure categories at the same monitoring sites based on calculated well activity.
Comparing a map of active well sites with data from air monitors located across the state, they looked for a correlation. For the six tested pollutants, the researchers found “fair to poor” agreement between the exposure categories determined using the well activity metric and actual measured levels.
The researchers see this as evidence that using well activity to categorize exposure to poor air quality was flawed, and call into question the results of the JHU studies linking disease outcomes to fracking, saying that further research was needed to determine the full impact of these misclassifications.
“The implication of these results was that using this methodology in epidemiology studies can result in significant exposure misclassification and therefore, uncertainty around reported risk estimates… Further study including disease outcomes and exposure classification of cases and controls would be needed to fully understand the nature and degree of the misclassification.”
The activist-driven studies claiming a connection between unconventional production and health risks have been used in several states to call for fracking bans. This study provides evidence to refute a key premise of these studies, asking whether exposure to airborne emissions can really be predicted by a person’s proximity to a production site.