*UPDATE* Six Flaws in a New Report Trying to Link Fracking to Infant Health Issues
UPDATE (12/13/17, 6:25 EST)
Earlier today, EID highlighted six facts to know about a new study claiming to show “evidence” that fracking is responsible for greater incidences of low birth weights, most of which call the study’s topline finding into question.
Fact No. 7 emerged soon after our blog post went live, and it reveals a probable motive for the researchers arriving at that inflammatory topline finding despite numerous limitations and contradictory data. Turns out one of the report’s major funders — the MacArthur Foundation — has provided Earthworks and Natural Resources Defense Council $10 million since 1978. Given that sort of financial commitment to a pair of groups that are actively seeking to ban fracking, it’s hard to imagine the MacArthur Foundation funded this study with hopes it would yield an objective, science-based conclusion clearing fracking.
After all, the MacArthur Foundation also gave anti-fossil fuel media outlet ProPublica an award in 2012 while noting, “ProPublica’s early reports on the dangers of natural gas drilling, or fracking, helped spark more critical attention and widespread debate on the issue.”
ProPublica, of course, has done numerous anti-fossil fuel stories, which can probably be traced to the fact that it has received tens of thousands of dollars from the Rockefeller Family Foundation over the years.
“Mining from oil shale is still more damaging, reducing prairies to barren moonscapes, contaminating surface and ground water, and emitting significant quantities of sulfur into the air. ‘Fracking’ for natural gas is one of the fastest growing means of fuel production. It contaminates ground water, reduces air quality, and leads to subsidence and seismic activity.”
Original Post (12/13/17)
A new study claiming to link infant health problems with oil and natural gas development is riddled with contradictory data and other flaws, raising questions about whether the authors oversold its findings to the press.
The study, authored by researchers from Princeton University, the University of Chicago and the University of California-Los Angeles (UCLA), claimed to show “evidence” that fracking has caused greater incidences of low birth weights in babies born to mothers living near Pennsylvania shale production sites. The press release declared that the study provides the “strongest large-scale evidence of a link between the pollution that stems from hydraulic fracturing activities and our health, specifically the health of babies.”
But the authors admit that the study was only exploring “potential” pollution, not “actual exposure,” and even said that such pathways are “not known with certainty.” Instead, the authors relied on a simple spatial correlation to suggest that the closer mothers live to production sites, the more likely they are to have babies with lower birth weights.
Yet their own data contradict that claim as well: according to the study, there was a greater incidence of low birth weights among individuals living farther away (3 km) from a well site than those living closer (2 km).
The authors even concede that their findings could have absolutely nothing to do with fracking. The study notes that the mothers they studied were “younger, less likely to have been married at the time of the birth, and less educated— characteristics that might lead to worse infant health outcomes even in the absence of fracturing.”
These aren’t small caveats. They’re flaws and limitations that directly contradict the inflammatory press release that the authors strategically circulated ahead of the study’s publication.
Here are six key facts to know about this study.
Fact #1. Study finds higher risk of low birth weight for mothers who reside 2-3 km away than 1-2 km — contradicting the report’s topline conclusion
The study purports to link proximity to well sites with lower birth weights. But as the following graphic from the third page of the study illustrates, the researchers found mothers who live 2-3 kilometers (km) away from oil and natural gas production sites are actually more likely to have babies with low birth weights than those who live 1-2 km away.
As the caption for this graphic explains,
“The left y axis of the graph indicates coefficients and confidence intervals (CIs) from a version of Eq. 1 in which ‘Near’ is replaced with 15 distance indicators representing the proximity of maternal residence to well sites; the coefficients represent the in utero effect on infant health of hydraulic fracturing (that is, when conception occurs after well spud date) at 1-km intervals from the well site… “
The chart clearly shows that the risk for low birth weight drops significantly in the 1-2 km bar (second from the left) and then sharply rises in the 2-3 km bar. There is no explanation as to why the data show mothers farther away from fracking sites in the 2-3 km distance category have a higher risk than those who are closer.
Interestingly, had the researchers used miles rather than kilometers, much of the data finding lower risk would have been included in a 1-mile-and-closer to fracking sites grouping and skewed the risk level down significantly. In other words, the odd choice of using kilometers is the only way the researchers can show a high result for mothers who reside near oil and gas development.
Fact #2. Authors link health problems to pollution, but admit they did not actually measure pollution
Like many studies that have attempted to link fracking to health problems, this study relies on spatial correlations and assumptions of exposure, rather than actual air quality data. As the authors acknowledge in the study:
“A limitation of our study is that given the nature of the available data, we are constrained to focus on potential exposure to pollution (which is determined by the mother’s residential location) rather than actual exposure that could be measured with personal monitoring devices.”
Despite the fact that the study’s conclusions are based on proximity rather than actual measurements of pollutants, co-author Michael Greenstone declared boldly in materials distributed to the press:
“This study provides the strongest large-scale evidence of a link between the pollution that stems from hydraulic fracturing activities and our health, specifically the health of babies.” (emphasis added)
In absence of actual measurements and proof of exposure pathways, the study cites a single deeply flawed study to support its assertion that emissions from oil and natural gas production sites exceed public health thresholds, while ignoring numerous studies that reached the opposite conclusion. From the report,
“One study measured various air pollutants weekly for a year surrounding the development of a newly fractured gas well and detected nonmethane hydrocarbons, methylene chloride (a toxic solvent), and polycyclic aromatic hydrocarbons, which have been shown to affect fetal outcomes.”
But the researchers neglect to mention that even the authors of that study — headed by Gasland star and noted fracking opponent Theo Colborn — were forced to concede that “the chemicals reported in this exploratory study cannot, however, be causally connected to natural gas operations.” Further, many of the measurements that they made showed emissions levels that were not high enough to trigger public health concerns.
EID’s recent Compendium of Studies Demonstrating the Safety and Health Benefits of Fracking details no fewer than five studies based on actual air measurements in Pennsylvania have found production emissions are protective of public health (one of these studies was conducted near a school). The Pennsylvania Department of Environmental Protection has also found that thanks to natural gas, emissions have been reduced by over 500 million tons.
Meanwhile, a recent health assessment from Colorado health officials collected over 10,000 air samples in parts of Colorado with “substantial” oil and gas operations, finding “low risk of harmful health effects from combined exposure to all substances during oil and gas development.”
Dr. Mike Van Dyke, head of environmental epidemiology at the Colorado Department of Public Health and Environment (CDPHE), noted when the data was released,
“This isn’t cherry-picked air sampling data. This is all air sampling data.”
CDPHE’s Chief Medical Officer and Executive Director Dr. Larry Wolk told the Greeley Tribune that “we don’t see anything to be concerned with” in regard to oil and natural gas development and public health.
Fact #3. Researchers admit sampled population would be expected to have similar health impacts even without nearby fracking
The researchers openly admit that factors other than fracking are potential contributors to low birth weight, including within their own sample set:
“The results show that using difference-in-difference reduces the potential for confounding fracking exposure with other determinants of infant health, but important differences in the evolution of marriage rates, race, education, and age remain. Although we control for all of the observable factors in our models, these differences suggest that there may also be unobserved differences across areas in other factors that influence infant health.”
“[M]others whose babies were potentially exposed to nearby fracturing in utero are younger, less likely to have been married at the time of the birth, and less educated— characteristics that might lead to worse infant health outcomes even in the absence of fracturing.”
There is also no indication in the paper that the researchers controlled for smoking or drug and alcohol use.
As maternal age increases, so do risks. That said, the study notes that in comparing siblings who were and were not exposed to fracking sites in utero, “exposed” siblings are almost always younger than the “unexposed,” indicating that the mothers in such instances were typically older at the time they had babies “exposed” to fracking.
Fact #4. Infant mortality rates in most heavily drilled Pa. counties declining faster than the rest of the state
The study’s press release not only asserts that fracking increases incidences of low birth weights, it also strongly insinuates that such instances are also driving up infant mortality rates,
“The research team found that infants born within a half a mile from a fracking site were 25 percent more likely to be born at low birth weights, leaving them at greater risk of infant mortality…”
But new research based on Pennsylvania Department of Health data refute this narrative, actually showing that infant mortality rates have declined in the most heavily drilled counties in the Marcellus since shale development began — even surpassing the declines seen in the state as a whole, as the following chart from that report shows.
Pennsylvania Department of Health data show that the six most heavily drilled Marcellus counties had greater infant mortality rate improvement than Pennsylvania as a whole from 2000 to 2014. This trend is even more notable, given the infant mortality rate in these six counties was higher than the overall Pennsylvania rate prior to the onset of shale development in 2000. By 2014, the infant mortality rate was lower in the six top Marcellus counties than for the state.
In is also worth noting that Pennsylvania Department of Health data show low birth weight rates were at or below the state’s average rates at the same time fertility rates increased from 2000 to 2014 in Pennsylvania’s six most heavily-drilled counties.
Fact #5. Researchers acknowledge small sample size
The study’s abstract touts the fact that the data is based on 1.1 million births from 2004 to 2013. Although that sounds like a large sample size, the researchers admit it is actually quite small, considering the study covers 10 years,
“A second limitation of our study is that even starting with the whole population of Pennsylvania births, we end up with a relatively small sample of children who were potentially exposed to fracking; this small effective sample size limits our ability to probe the shape of the distance-exposure relationship and also limits our ability to obtain precise estimates from models with mother fixed effects.”
This is especially true with regard to the sample size of mothers living 1 km or closer to fracking sites, which are much smaller than the groups farther away from well sites, making them more susceptible to anomalies and outliers. As this graphic from the study shows, the bar graph to the far left indicates fewer than 1,000 births from mothers who live 1 km or closer to fracking sites were included in the study. Again, considering study spans 10 years, this is an extremely small sample size.
Much like the aforementioned data — which at best, is inconsistent and at worst is flat out inaccurate — this small sample size issue further weakens the study’s topline conclusion.
Fact #6. Report lauds fracking’s benefits, which authors have also previously touted
To their credit, the authors do acknowledge the benefits of the shale revolution, including reducing energy prices, decreasing pollution from power plants and bolstering U.S. energy security. From the study,
“The growth in unconventional gas production involving hydraulic fracturing (“fracking”) has transformed the energy landscape, reducing energy prices, decreasing conventional air pollution by displacing coal in electricity generation, disrupting international energy trading arrangements, and increasing the prospects for energy self-sufficiency for the United States.”
These positive caveats aren’t entirely surprising considering two of the study’s authors have made positive comments in the past regarding fracking’s health and economic benefits.
Greenstone has previously said,
“There’s a strong case that people in the U.S. are already leading longer lives as a consequence of the fracking revolution.”
He has also previously said that there have been “tremendous health gains” due to increased natural gas use for electricity generation and that, “honestly, the environmentalists need to hear it.”
Co-author Janet M. Currie, a Princeton University economics professor, also co-authored a study with Greenstone last year that found local communities have enjoyed “enormous benefits” from fracking. The study found average annual benefits of $1,200 to $1,900 per household for those who live near fracking sites, as well as a six percent increase in income and a six percent increase in housing prices.
Greenstone’s prior statement that, “There’s a strong case that people in the U.S. are already leading longer lives as a consequence of the fracking revolution” is still valid. It also makes the bold declarations in this latest study more puzzling, especially given the flaws and contradictions in the data used to support such claims.
Notably, data from groups like UNICEF have shown that widespread air pollution in developing countries is contributing to low birth weight and a myriad of other childhood health issues. But fortunately here in the U.S. – thanks largely to increased use of clean burning natural gas made possible by the shale revolution – we have the fewest air pollution-related deaths per capita in the world, and we are dramatically reducing emissions such as fine particulate matter that have been linked to birth defects and other serious health issues.
We have achieved these reductions at the same time as we’ve emerged as the number one oil and natural gas producer in the world, helping to lift our economy out of the depths of the Great Recession to boot.
If the questionable conclusions of studies such as this one are mainstreamed into the public consciousness even further, we stand to miss out on the substantial health and economic net benefits of fracking that these researchers have previously highlighted.
NOTE: Information regarding a flaw regarding calculation errors, which still may be valid, has been moved pending additional review.