More on the UPitt Health Studies

Uni Blake
Posted September 1, 2023

Everyone wants to see communities protected from potential impacts of industrial operations – natural gas and oil companies included. Safety and environmental protection are integral to sustainable energy development that benefits families and businesses while also reducing emissions.

Our industry welcomes public health research that helps improve our operations. Needed are reliable data and analysis that stand up to the rigors of established scientific study.

Using that metric, three recent environmental health studies from the University of Pittsburgh’s (UPitt) School of Public Health have limitations in helping us know whether unconventional oil and natural gas development (typically using hydraulic fracturing) can be linked to specific health concerns in nearby communities.

As a health scientist trained in understanding exposure as it relates to environmental and public health, my concerns with the UPitt studies fall into two main categories:

1. Incomplete data

All three studies relied on existing data and medical records that were not originally collected for research purposes.

As a result, important “confounders” are not included. Confounders are hidden factors that can also explain the connection or association between two things, making it harder to understand true cause-and-effect. Think of the association between ice cream and sunburn: Does the ice cream cause the sunburn? Of course not, but people who eat ice cream tend to spend more time in the sun, which directly causes sunburn. So, exposure to sun “confounds” the relationship between ice cream and sunburn.

I say all this not to trivialize the study findings or the concerns they initiate but to provide context as to why it is important to understand the potential role of confounders. Administrative datasets used in the UPitt studies could have missed important confounders, and those that were included may have had missing or incomplete information.

In short, we believe that the connection or association between a health outcome and proximity to a natural gas well could change if we had more information about other factors that might be influencing the situation.

2. “Proximity” vs. “exposure”

As I noted in this blog before the studies were issued, there are limitations in relying solely on “proximity” as a surrogate for “exposure,” which refers to the things we are examining to see if they could affect our health.

This could include factors like changes in air or water quality, as well as stress. However, because of the study design, it is difficult to say that a particular exposure (changes to air or water quality) directly caused the health outcome being studied.

These studies only look for connections and relationships, but to say one thing causes another needs more carefully thought-out research.

For example, in UPitt’s childhood cancer study, proximity or nearness to both unconventional oil and natural gas development and conventional oil and natural gas development was determined using a child’s birth certificate. This was then carried forward for the entire “exposure period” between birth and diagnosis, which could have been as long as 29 years. A lot can happen over the course of nearly three decades.

While proximity can indicate potential exposure to unconventional oil and natural gas development, it doesn't account for variations in individual behaviors, such as time spent outdoors, ventilation or protective measures that may have been taken.

Proximity-exposure relationships also can vary widely depending on factors such as wind patterns, barriers and geographical features, making proximity an imperfect proxy for exposure levels

Recognizing these limitations and the concerns expressed by community members, our industry remains committed to transparent communication and ongoing research. We want to provide a comprehensive perspective that factors in the complexity of interpreting environmental health-related concerns, especially multi-decade studies.

The outcomes of this and other studies continue to drive meaningful discussions and scientific inquiry. We view this as an ongoing dialogue rather than a definitive endpoint.

In the coming weeks, we will analyze the reports to gain a more nuanced understanding of the data and its precision and implications. We will continue collaborating with community members and subject matter experts to evaluate study designs and contribute our perspective to support a science-driven understanding of environmental health concerns.