What are the key findings and conclusions of the University Network for Human Rights household health survey?

The University Network for Human Rights recently released the results of a health survey of over 500 households in the area near the Denka/DuPont plant, which emits chloroprene at dangerous levels according to the EPA. The University Network study—which The Guardian has called “the most detailed and comprehensive evidence to date that [residents] are at an especially pronounced risk of cancer and other negative health effects due to toxic chemicals in the air”—found that the cancer rate among survey respondents who live within a mile of the plant is nearly double the expected rate (12.4% versus 6.9%).

In the study, we assessed ten different chloroprene-associated health conditions in both adults and children, and every single condition was more prevalent in households within a mile of the plant than in households farther away. Our findings—that cancer and illness levels among those surveyed are unusually high and correlated with distance from the plant—are deeply disturbing and underscore the need for additional health studies in the area near the Denka/DuPont plant and throughout the industrial corridor. Local authorities must act immediately to compel Denka Performance Elastomer to reduce chloroprene emissions to EPA-recommended levels.

What was the study methodology?

Survey implementers used a randomized list of household addresses to survey one person per household. The person surveyed (the “respondent”) was asked to provide health information about themself as well as all other household members.

Survey-based approaches, including those that rely on one household member to provide health information about others who live in the same household, are well-established in epidemiology. Such methods are known as “shoe-leather epidemiology” because data is collected by going door-to-door (and wearing out shoes) in affected communities. When an urgent public health problem arises, shoe-leather epidemiological studies are a common first line of response.

Survey respondents were not asked to provide medical records. There are major legal and ethical implications involved in this, and asking survey respondents for medical records is not the norm or established practice for these types of studies.

The 14 Stanford University undergraduates who implemented the survey underwent intensive training by Stanford experts in survey implementation best practices and techniques, including a practical training component that involved mock implementation of the survey with actors as respondents. The survey was implemented in a rigorous, consistent, and systematic manner. 

Why weren’t the study’s analyses of cancer prevalence broken down by specific types of cancer?

Breaking down our analyses by cancer type would have been a flawed approach for several reasons. Disclosing small numbers of people in a concentrated area who have reported specific, rare cancer types implicates serious privacy and confidentiality concerns, as well as statistical concerns having to do with sample size. In addition, although we know that chloroprene is a likely human carcinogen, there is a lot that we do not know about the specific types of cancer linked to chloroprene exposure. Given these issues, focusing on only certain types of cancer would not have been scientifically sound.

How do the study data compare with data published by the Louisiana Tumor Registry?

Denka continues to invoke the Louisiana Tumor Registry (LTR) to defend its claims that cancer rates are no higher in the area near the plant than elsewhere in the state. Comparison of the survey data with LTR data (now published by census tract) is not scientifically sound. Census tracts are not homogenous; some homes in census tract 708, for example, are close to the Denka/DuPont plant while others are relatively far from the plant. Thus, LTR data for census tract 708 simply cannot be compared to the results of our study, which focused specifically on the area immediately surrounding the plant. In July 2019, moreover, the LTR published updated guidance clearly stating that the link between cancer and pollution cannot be established using LTR data alone. In addition, unlike the LTR, our health survey collected data not only on cancer, but on a range of other health conditions as well.

Is the study peer-reviewed and published in a scientific journal?  

The study, which was designed and implemented with input and guidance from statisticians and epidemiologists at Stanford University, has not yet been peer-reviewed and published in a scientific journal. Although publication in a peer-reviewed journal may satisfy some academics who are untouched by pollution in their communities, pursuing publication would have significantly delayed the release of our findings. Given the urgency of the situation that residents are facing in St. John Parish, it would have been irresponsible for us to withhold this information. This does not, however, foreclose future publication in a peer-reviewed journal, and we are pursuing that avenue as we move forward.

For more information, please contact: