Three weeks ago, University of Colorado researchers published a study linking diagnoses of a certain kind of blood cancer to the likelihood of living within 10 miles of oil and gas operations.
Last week, the Colorado Department of Public Health and Environment (CDPHE) published its own report, a two-part review of emissions data and existing studies related to the health impacts of oil and gas. Its conclusion? The risks are low.
Two reports, two very different sets of findings. The methods of the study and review were quite different, so they can’t be directly compared, but both asked the same basic question: Are oil and gas developments harmful to human health? The answer, as both sets of researchers concluded, is that more study is needed. But the timing of the two reports and CDPHE’s contrary findings have reinforced the belief among some Colorado environmentalists that CDPHE is minimizing potential health risks posed by oil and gas development.
The reaction comes amid increasing tensions between Colorado residents and the oil and gas industry. Homeowners continue to attempt to gain control over fracking in their neighborhoods, only to come up time and again against state law, which preempts local governments from limiting or banning residential drilling.
And President Donald Trump’s moves to remove federal methane regulatory requirements, slash the budget for the Environmental Protection Agency and cut environmental regulations on air and water — have heightened the feelings of alarm. Even before Trump took office, news broke that the EPA downplayed fracking’s risk to drinking water.
Last September, The Colorado Independent interviewed Colorado’s chief medical officer, Dr. Larry Wolk, after he said he didn’t “see anything to be concerned with” when it comes to the impact of fracking on public health.
Dr. Wolk declined requests for a follow-up interview, but CDPHE spokesman Mark Salley recommended The Independent speak with agency epidemiologist Mike Van Dyke instead about the review and its findings. The review came at the recommendation of Gov. John Hickenlooper’s oil and gas task force.
Van Dyke addressed current and upcoming research into the health impacts of fracking, the limitations of different research methodologies and the concern among some Coloradans that the health department is dragging its feet in acknowledging a link between oil and gas development and public health problems.
The Colorado Independent: Can you tell our readers a bit about your background, and what kind of work you do for CDPHE?
Mike Van Dyke: I am currently the chief of environmental epidemiology, occupational health and toxicology at the CDPHE. I have a PhD in environmental health from Colorado State University with a focus in exposure science, and I have about 20 years working with occupational and environmental exposures and health effects that might be related to those.
CI: I want to discuss two studies released in February, one led by CU researcher Lisa McKenzie and one by the CDPHE, about the health impacts of fracking. It’s clear that more research is needed, but McKenzie’s study found a link between oil and gas developments and diagnoses of acute lymphocytic leukemia in children ages 5-24, and the CDPHE study found that “the risk of harmful health effects is low for Coloradans living near oil and gas operations.”
I have heard from several community members who have called the CDPHE study a response to, or even a rebuttal against, McKenzie’s research. Any relationship between the two, timing-wise?
MVD: No, absolutely not. Our study had been in the works for over a year. I think if anything, we saw that we were almost done and we saw a good deadline with the Broomfield presentation [a community forum designed to educate Broomfield City Council and residents about oil and gas] coming up, which we thought was a good opportunity to present that information. So we pushed to get to that deadline, but it had nothing to do with the McKenzie study.
CI: Let’s talk about McKenzie’s study first. What is your reaction to her findings, which include that patients aged 5–24 diagnosed with acute lymphocytic leukemia were 4.3 times as likely as patients diagnosed with non-blood cancers to live within 10 miles of oil and gas developments?
MVD: First, I want to say that these are the kinds of studies that need to be done. These are important studies. As a department, we view all of these types of studies.
But this is also kind of a base level environmental epidemiology study. It’s registry-based, meaning that they used administrative data from the [Colorado Central Cancer Registry], and they really used location as a surrogate for exposure, and they had a significant finding. There are lots of studies out there like that, studies for living near a highway, living near a gas station, and this is really similar to all of those things. These studies are, of course, important, but what is more important is that we really put this study in context of the whole body of science surrounding this issue.
[McKenzie] does a really good job of outlining the limitations of the study. The conclusion was yes, this is a significant finding, but a lot more work needs to be done. So this is just research that suggests more research needs to be done, not research that definitively links oil and gas exposure to cancers in this age group.
CI: But do you really think that’s all it suggests — that more research needs to be done? Does it not also suggest a relationship between oil and gas development and certain childhood cancers?
I think it suggests that there might be, but I think what it really is, is there are a lot of alternative explanations that could be proposed to explain this same relationship.
CI: Like what?
MVD: It could also be due to living near a road, there could also be unmeasured factors in terms of things like agricultural exposures, there could be family exposures like people smoking indoors. None of those things were really accounted for in the design of the study, nor should they have been. That kind of research requires a much more detailed and expensive study. I don’t want this to come across as if I’m bashing this study, but I do want to come across as saying that this study has significant limitations.
CI: In her discussion of her findings, McKenzie wrote that the “use of cancer-controls, the limited number of [cancer] cases, and aggregation of ages into five year ranges, may have biased our associations toward the null.” Pardon my ignorance — is she saying that those factors would have biased the study towards the conclusion that oil and gas has more or less of an effect on health?
MCD: That would mean biased more towards no effect.
CI: So she’s saying that the likely bias would be towards a muting of the effects — does that change your interpretation of the study?
MVD: There’s still a huge number of limitations, so no, it doesn’t change. I think in the bigger picture of things, what we do at the health department with these sorts of issues — and we do this with studies related to oil and gas, marijuana and other things — is we put things into a kind of systematic review framework. And that’s what we did in our report: You really try to put all the findings on the table, and you rate the findings with, “Is this a high, medium or low quality study?” and that’s based on the strengths and limitations of the study. We try to do that for every study, and then we put the findings together and we have at least an answer that is understandable to us, and we can write about it, and say what is the quality and quantity of scientific evidence that supports this linkage?
What it comes out to is you have Lisa McKenzie’s study, and it did show an association between living near oil and gas and cancer cases in that age range. But there is another study — and the other study is probably lower quality than Lisa’s study, but there is another study — and it didn’t show an association between those things.
Where we are now with childhood cancer, or should I say with leukemia and hematological types of cancer, is that we have mixed evidence as to whether there’s an association between these two. And when you have mixed evidence, when you have limited evidence, that says you need to research this a lot more.
So it really needs more science to confirm or dispute it. And, if you look at some of the studies that are out there, there have been studies that show increased leukemia in kids who live near gas stations, but later studies did not find the same thing. Later studies show that that finding was just in that one study, which tells you the result was likely due to some issue in that study.
I’m not saying whether or not that’s going to be the case with Lisa McKenzie’s study, but that’s common. That’s what happens in science.
CI: Let’s move on to the CDPHE study. One of the study’s main research questions was: “Do substances emitted into the air from oil and gas operations result in exposures to Coloradans living near oil and gas operations at levels that may be harmful to their health?”
To my understanding, that question is focused on looking at how current exposure levels compare to agreed-upon safe levels, not at whether those exposure levels are actually safe. As I understand it, it’s essentially saying that X amount of benzene is considered safe, rather than testing whether there is a correlation between benzene levels and certain health problems. What are the strengths and weaknesses of such a question?
MVD: Our current report has two pieces. The first part is essentially all the air samples that we could find that were taken in areas near oil and gas, and we really, in a conservative way, took the max concentration and the max average concentration from those datasets, and compared them to what would be called “safe” levels by the U.S. EPA, or by other states if the EPA did not have a value.
Those values are really based on all the studies that have been done on those particular chemicals, and safety factors have been applied to the levels in those studies. The way people interpret those results are that they are conservative estimates to protect nearly all people from health effects.
What we found was that based on these data, there were no chemicals or substances that exceeded those safe levels. This was a pretty conservative way of doing things, and then we added them together, and even with all the chemicals combined, you barely got above the safe exposure level.
But there are limited methods for really doing a risk assessment on multiple chemicals, and a conservative way of doing things is saying, ‘Let’s group the chemicals by their health effects.’ So if you have three chemicals that cause asthma, you divide the amount of each of those by their safe levels, and then add those three together. What you want is for those to be below one — 100 percent. The problem with that in this kind of analysis is that people aren’t usually doing risk assessments with 60 chemicals; they’re typically doing risk assessments with 3 or 4.
At some point, just the sheer number of chemicals you’re including in the analysis puts you above one.
CI: Wait…so if you’re saying that the sheer number of chemicals can put you above 100 percent safe levels, does that make oil and gas activity, if it includes such a large number of chemicals, inherently dangerous?
MVD: When we identified the chemicals we used in our risk assessment we included 62 chemicals, and those are all probably in some way related to an oil and gas site. However, you can’t attribute the entire concentration that you measure to an oil and gas site. There’s going to be a highway nearby, there’s going to be a gas station. The truth of all of this is that all of these chemicals are incredibly common, and most of them you can measure in air anywhere.
CI: Can you give me an idea about roughly what percentage you can attribute to oil and gas? Is it half?
MVD: The best I can say is that you can attribute some, and it’s going to depend on the chemical. It’s really hard to say, when you have a very common chemical, where it’s coming from. The CSU study that they did in the Front Range and Garfield County [a multi-year study “aimed at characterizing the extent of air emissions from natural gas extraction activities”], those results were released last summer. The strength of the CSU study is that that is the only data for which we are going to be able to say, ‘This chemical came from an oil and gas site.’ It’s a lot stronger dataset than just having these ambient air measurements.
CI: The CDPHE study found that “Cancer risks for all substances were within the “acceptable risk” range established by the U.S. EPA,” and that, quote, “Although well within the acceptable risk range for cancer and non-cancer effects, benzene, acetaldehyde and formaldehyde had the highest estimated risk levels and are high priority for continued monitoring.”
I think many community members might balk upon realizing that the CDPHE is not choosing to further limit, through, for example, a change in policy towards oil and gas development, substances which are a high priority for continued monitoring. What would you say to these residents to ease these fears, and fears that the CDPHE is choosing to delay action despite what appears to be a link between oil and gas and health concerns?
MVD: That question gets away from the science, and I answer questions based on the science. What I can say, from a science perspective, is that science doesn’t support the supposition that all exposure is harmful. Science supports that there is a safe level of exposure for most of these things.
CI: So what would research look like, what amount of data would be required to make the CDPHE conclude that oil and gas is harmful to public health?
MVD: We’d have to have a lot better exposure data, first of all. What we use for exposure data definitely has limitations, and that’s the direction we’re headed in using the CSU data. Basically what we know from the CSU study is we know how much of each of these chemicals is emitted from each source. In order to translate that into, “What does exposure look like for somebody standing 1000 feet away from a well?” it has to go through a computer model to do that. That’s what we’re going to get for the study we’ll release in 2018.
It’s not necessarily the number of studies, but you need a high quality epidemiology study, and unfortunately those typically involve a whole lot of time and a whole lot of money. Epidemiology is expensive and time consuming. The data that’s used is mostly administrative data. When you could look at the birth outcome studies, you could look at some of the cancer studies, the data you have in those studies is the address at diagnosis or the address at birth. With that data, you’re making the assumption that people have lived there their entire life.
It’s even worse for a cancer study, because you’re talking about many years. With a birth study, it may not be as bad, but for a cancer study, you’re talking 24 years [McKenzie’s study looked at cancer cases in people aged 0-24]. What you need is the full address history of a person. You don’t know, for example, if a person used to live across the street from another industrial operation and moved to this place a year ago. That’s an extreme example, but you have no idea the real exposure history of these people. So you need a full address history, and doing that becomes much more time consuming and expensive.
All of this makes these studies very hard to interpret, so we need a better study, and unfortunately that better study is a lot more expensive.
CI: Is that something we can expect going forward? Why hasn’t the CDPHE begun such a study?
MVP: It’s unusual, first of all, for a state to pay for a risk assessment like we’re doing with the CSU data. That doesn’t happen very often, and this is a good thing.
In terms of an epidemiology study, I am not aware of any state government — and I’m thinking pretty hard here — that has funded or undertaken one. State governments don’t fund multi-million dollar epidemiology studies anywhere, for anything. Those really expensive epidemiology studies are normally done by, funded by, the federal government.
CI: The CDPHE study looked at the results of a variety of studies and compared the results. How did you make sure the studies you included in your research weren’t biased?
MVD: We didn’t really, that wasn’t part of our inclusion or exclusion criteria. There’s not a lot of studies out there on this issue. In fact, there are about 12, and we used them all. I would say that these were published studies, these were not industry reports, so I would say that they’re likely unbiased.
CI: Again, what can you say to Coloradans who have reacted to this research with concerns that there’s a reluctance within the CDPHE to accept a link between fracking and public health problems?
MVD: I can tell you, as clearly as I can, that there is not an agenda. We look at the science, we evaluate the science, and the timing [with the McKenzie study] was coincidental, so the two things have nothing to do with each other. I think that what I could say about Colorado is that we are in a situation now where we have a program called the Oil and Gas Health Information and Response Program. This program is funded by state dollars, and its sole focus is, one, to respond to citizens’ health concerns, and the second is to review the science around oil and gas and health. So we have people for whom that’s their job. They don’t work in regulations, they work in the health side of the health department. They’re scientists. And that’s a good situation, and that’s not a situation that we’ve always had in Colorado.
CI: What kind of calls do you get, and what kind of action do you take?
MVD: We see people calling in about all kinds of things. The best resource to really see those results is in the legislative report that is listed on the website.
What we have now is a centralized place for health concerns across the state. We want people to call us and tell us about their health concerns. We may not be able to solve their health concerns, but knowing about their health concerns is important.
This interview has been edited for clarity and brevity.
Photo credit: Ted Wood, The Story Group