NPR Uses Junk Science in ‘Fracking Harms Health’ Claim
As part of its woefully misleading and one-sided series attacking the oil and gas industry, NPR’s Marketplace published a story on Wednesday suggesting fracking chemicals are harming public health. In an effort to support that narrative, reporter Scott Tong lists a dozen studies he (presumably) feels best support the argument, while also adding the following disclaimer:
“While no medical diagnoses have been revealed to be caused directly by these oil and gas drilling chemicals, cause and effect can be difficult to prove.”
Translation: None of the studies listed by MarketPlace prove causation, which is a systemic shortcoming of literature attempting to link fracking to health problems that have been flagged by environmental research group Resources For the Future (RFF). But this is just one of a myriad of flaws shared by each of the studies included Marketplace’s list. Let’s review.
Study #1: McKenzie et al., 2012 (link)
Marketplace leads off with this infamous study, claiming it shows “air pollution caused by fracking may lead to health problems for those who live near natural gas drilling sites.” There is no mention at all that this study is so seriously flawed that the research team responsible for the paper was fired by the Colorado county that commissioned it almost a year before the study was published. Garfield County environmental health chief Jim Rada also disavowed the paper for its “significant” data limitations.
Though the study’s conclusions were based on actual air samples (a rarity among such literature), it exaggerated emissions from well development by at least 10 times, due largely to the fact that the researchers failed to into account exhaust fumes from a major interstate highway less than a mile away.
The study’s long list of flaws also includes the fact that it included no background emissions data and also failed to note the cancer risk detected was not above the national average. The researchers’ methodology was so bad that the Colorado Department of Public Health and Environment (CDPHE) reneged on its planned funding of the research. CDPHE has been a vocal critic of lead researcher Lisa McKenzie’s work ever since, which brings us to the next study on Marketplace’s list.
Study #2: McKenzie et al., 2014 (link)
Marketplace claims this equally infamous McKenzie-led effort finds “an association between those who lived within a 10-mile radius of (oil and gas development) and congenital heart defects and possibly neural tube defects.”
Marketplace fails to note that McKenzie and her research team ignored a host of other risk factors, including smoking during pregnancy, access to prenatal care and genetic history. This study was so flawed that CDPHE issued a statement prior to the papers’ release warning the public not to be misled by its conclusions. Dr. Larry Wolk — the CDPHE’s Chief Medical Officer and Executive Director, and recent U.S. Environmental Protection Agency advisory board appointee — said the researchers used “minuscule” statistical differences to claim some kind of connection, a connection which public health officials rejected.
“[W]e disagree with many of the specific associations … [and] a reader of the study could easily be misled to become overly concerned,” Wolk said in a January 2014 statement.
In the wake of this rebuke, one of the study’s co-authors admitted the paper’s shortcomings, telling Rocky Mountain PBS, “It’s certainly not a conclusive study, and it doesn’t demonstrate that pollutants related to shale development have caused birth defects.”
Study #3: Jackson et al., 2013 (link)
Marketplace notes this Duke University study based on water well samples in Pennsylvania “found elevated levels of methane, ethane and propane gases,” with the clear inference being that the study proves fracking contaminates groundwater.
But Marketplace fails to note that researchers found methane in virtually every well they sampled, regardless of whether the wells were close to shale production sites. In fact, more than 50 of the wells among the 141 total homes sampled were located nowhere near natural gas wells. Among the researchers’ samples, drinking water supplies that registered measurable methane concentrations were almost equally likely to be located either close to a natural gas well, or not close.
Naturally occurring thermogenic methane has been widely present in Pennsylvania water wells for decades. This fact was confirmed by a USGS study that found thermogenic methane pre-dating Marcellus Shale drilling in wells located in the same areas the Duke study included. The Duke study also found no evidence of fracking fluid in the wells, but these facts were nowhere to be found the Marketplace story.
Study #4: Casey et al., 2016 (link)
Marketplace reports that this examination of 10,000 births between 2009 and 2013 in north and central Pennsylvania found “mothers living in the most active fracking areas were 40 percent more likely to give birth prematurely, and 30 percent more likely to have their pregnancy labeled high-risk.”
EID pointed out soon after the study was released that the premature birth rate recorded in the areas closest to where wells exist is below the national premature birth rate. The researchers also failed to use available baseline data, take measurements and didn’t factor in genetics and socioeconomic factors. No wonder the study drew sharp criticism from Dr. Tony Cox, a clinical professor of biostatistics and informatics at the University of Colorado-Boulder. RFF also discussed many of the flaws flagged by Cox and EID in a recent report, most notably the fact that, “The authors found no correlation of unconventional natural gas activity with Apgar score, SGA or term birth weight.”
Study #5: Jemielita et al., 2015 (link)
Marketplace notes that this University of Pennsylvania and Columbia University study “found that those who live near (oil and gas) activity have a higher likelihood of being hospitalized for cardiac, neurological, urological, cancer-related and skin-related problems.”
But as EID highlighted when the study was released, the county in the study area with the highest number of wells (Bradford) actually had the lowest overall inpatient occurrences, which runs counter the topline conclusion touted by Marketplace (and general, good ole’ fashion common sense). RFF also criticized the study for failure to prove causation and failure to determine other contributing factors.
Study #6: Stacy et al., 2015 (link)
Marketplace reports that this study found mothers who “who lived closest to a high density of fracking wells in Pennsylvania were 34 percent more likely to give birth to infants who were small for their gestational age.”
As EID noted when the study came out, the paper (as usual) failed to prove causation and included contradictory data. For instance, none of the average birthweights found were actually considered “low” by the medical definition of the term, and all regions studied had average birthweights above the national average. Furthermore, the study’s data actually showed that the average birth weight in the study area farthest away from shale gas wells was 3,343.9 grams, much lower than those in the second region (3,370.4) and the third region (3,345.4), which were closer to shale gas wells. RFF took note of the latter contradictory data:
“The study did not find any significant effects of well density on premature births, except for a higher average birth weight and a lower share of premature infants born to mothers living in the second exposure quartile — an odd result…”
Marketplace also failed to report this study was funded by the anti-fracking Heinz-Endowment. But hey, at least they’re consistent (see EID’s debunk of Marketplace’s first story in its oil and gas hit piece series for more details on its disclosure issues).
Study #7: Macey et al., 2014 (link)
Marketplace reports that this study found eight “volatile chemicals” near wells in five states, noting “benzene, a carcinogen, and formaldehyde were among the most common.”
There is no mention of the fact that this infamous study was spearheaded by anti-fracking Global Community Monitor (GCM), which has affiliates around the country known as “Bucket Brigades” — groups of activists that literally collect air samples in buckets lined with plastic bags to suggest that air quality is being impaired by oil and gas development. In the study, the researchers allege that “potentially dangerous” air pollution is “frequently present near oil and gas production sites” on the basis of those “bucket” tests.
Not shockingly, considering the sampling technique used, the study’s results have been roundly debunked for years, and only seem to crop up in anti-fracking activist commissioned reports and sympathetic media stories (such as this one). To be perfectly clear, GCM has admitted, “The Bucket Brigade is not a scientific experiment. Our focus is on organizing. We use science, but only in the service of organizing.” Translation: GCM prefers advocacy over science. And any serious journalist should at least disclose that fact before presenting its findings as objective science.
Study #8: Rabinowitz et al., 2015 (link)
Marketplace reports this study found Washington County, Pa., residents who live close to natural gas facilities are “more likely” to report skin issues, headaches, and nosebleeds and have twice the number of health issues as folks who live at least two kilometers from natural gas sites.
But there is no mention of the report’s laundry list of major flaws. First off, the lead author emphasized that the study failed to prove causation and was “preliminary.” RFF also noted in its aforementioned report that the study was based on self-reported symptoms, a data collection method potentially “subject to more bias” than alternate techniques. And considering an anti-fracking group, Southwest Pennsylvania Environmental Health Project, was responsible for recruiting participants for a study funded by the anti-fracking Heinz Foundation, it’s difficult to imagine there wasn’t a whole lot of bias on many, many levels.
Study #9: Rasmussen et al., 2016 (link)
Without scrutiny, Marketplace — as many media outlets did before it — cites this well known Johns Hopkins Bloomberg School of Public Health study as evidence that there is a connection between “worsening asthma symptoms” and proximity to fracking.
But aside from the fact that this analysis of electronic health record data fails to prove causation (we know, it’s starting to sound like a broken record), it is also essential to understand that a vast majority of the more than 35,000 asthma patients whose medical records were the basis of the study lived in areas with little to no shale gas development. Furthermore, the data revealed that asthma attack rates were actually much higher in numerous counties with no shale development than in heavily drilled counties. And oddly, heavily-drilled Washington County wasn’t included in the report. Notably, lead researcher Brian Schwartz is a fellow at the Post Carbon Institute and the study was partially funded by the anti-fracking Robert Wood Johnson Foundation. RFF also noted that because “the characteristics of populations within each exposure quartile” were not evaluated that it is “difficult to assess whether the results are credible.”
Study #10: Kassotis et al., 2015 (link)
Marketplace runs this study out as evidence that fracking can harm the reproductive health of men, reporting, “… one study exposed a set of male mouse offspring to a mixture of chemicals that represented what humans would likely be exposed to from wastewater and drinking water” and that these mice had “a lower sperm count, higher testosterone levels in the blood and larger testicles in adulthood.”
While the latter fact could definitely be filed under the “too much information” category, it’s the info Marketplace left out that’s more important: These researchers actually created a cocktail of 24 chemicals that, if consumed in very large doses, are known to cause health problems. They then injected this cocktail – in very large doses – into the water of pregnant mice and declared that it caused the baby male mice to have low sperm counts.
Lead researcher, Christopher Kassotis perhaps put it best when he admitted, it’s “unlikely people would ever be exposed to doses quite as high.”
Study #11: Bamberger et al., 2009 (link)
Marketplace goes back nearly a decade for this gem, reporting the study found “Sixteen cattle died after drinking a ‘mysterious fluid’ adjacent to a natural gas drilling rig.”
The researchers responsible for this study have fully admitted that they didn’t produce a scientific assessment at all, explaining, “By the standards of a controlled experiment, this is an imperfect study.” Dr. Ian Rae, a co-chair of the Chemicals Technical Options Committee for the U.N. Environment Programme, explained this major flaw, “It certainly does not qualify as a scientific paper but is, rather, an advocacy piece that does not involve deep…analysis of the data gathered to support its case.”
Study #12: Levinstein et al., 2013 (link)
Marketplace reports that this study shows, “Fracking could be having a negative impact on the dairy industry” and that the study found “the most heavily drilled areas saw a 30 percent loss of milk cows.”
If you’re wondering why this is included in a list of studies meant to argue that fracking harms health, you’re not alone — this study isn’t even about health. Nonetheless, the reality is dairy farming in Pennsylvania has been declining since long before shale development began. More importantly, recent research from the Endless Mountains Heritage Region actually shows that farms overall have flourished since shale development began.
Marketplace wrong on EPA study, too
All told, Marketplace makes reference to just one reputable study in its effort to support its narrative that fracking causes health problems — EPA’s landmark five-year groundwater study. But predictably, it mischaracterizes its conclusions as well:
“The Environmental Protection Agency conducted a five-year scientific study of fracking’s effect on U.S. drinking water. … While one released version of the report downplayed this connection — stating that fracking has not led to “widespread, systemic impacts” — the EPA later reversed that conclusion.”
Simply stated: EPA did not reverse course on its conclusion that fracking has not led to “widespread, systemic” impacts. The data in the EPA’s final report was the same as what appeared in the draft version. Former EPA Deputy Assistant Administrator Thomas Burke even clarified that the number of instances in which groundwater was affected was “small.” In a recent CBS This Morning interview he repeated that that finding, noting, “The overall incidence of impacts is low.” EPA reached that conclusion despite the fact that it expanded the definition of fracking to include just about every oil and gas activity even remotely associated with the process.
Conclusion
Marketplace’s one-sided take isn’t all that surprising considering it received more than $2.1 million from the anti-fracking Tides Foundation to launch “Sustainability Desk” program responsible for this week’s series.
But the collection of flawed anti-fracking research listed in this story actually runs counter to the article’s obvious objective — presenting convincing scientific evidence that fracking harms health. Instead, it shows exactly why RFF’s recent review of the most prominent fracking health studies noted that “all had shortcomings that were most often significant.” RFF also reported “contradictory results for each impact” and noted that studies on birth defects, hospitalizations and multiple symptoms, in particular, were cumulatively deemed to be of “low quality.”
At the end of the day, the actual data from these studies do not justify the scary headlines they are producing, and this Marketplace list is perhaps the most comprehensive example of that fact to date.
Read more at EID
Trackback from your site.