Locate a Member
Enter the first part of a member's last name to search
New Member Registration
Search this Site
12635 E. Montview Blvd., Suite 270
Aurora, CO 80045
P: (720) 859-4149
F: (720) 859-4158
Allergic Disease in Neighbors of Concentrated Animal Feeding Operations
Investigator: Leland Stillman
Mentor: Larry Borish, MD
Summary: Concentrated animals feeding operations (CAFOs) generate significant air and water pollution to surrounding communities. Air pollutants include carbon dioxide, hydrogen sulfide, methane, volatile organic compounds, and particulate matter. These pollutants have been reported to cause significant respiratory symptoms in humans, and specifically in CAFO workers. A case-control mail survey was conducted using address based sampling (ABS) in a region of North Carolina with a high density of CAFOs regarding frequency and severity of asthma symptoms. The response rate was less than 1%, and the results were not significant.
Course of Research: The original research proposal was modified twice times in the course of the research project, both times due to constraints of the IRB.
The original proposal to reenroll participants in a previous UNC Chapel Hill epidemiological study on air pollution were not possible, due to the fact that the IRB those participants had signed did not allow for them to be re-contacted by those in charge of the study. This led to a search for an alternative method of study. Address based sampling was chosen, given that the University of Virginia Center for Survey Research has expertise in conducting studies of this kind, and that this study design, unlike the previous study design, would use a random sampling, eliminating a substantial source of bias. This has been the first randomized survey study of the health effects of concentrated animal feeding operations.
Following this, it became clear that the IRB at the University of Virginia would not approve the study without substantial changes in our methods. The IRB refused to expedite approval of the study, despite its non-invasive nature, unless skin prick testing was not included. As this would have led to unacceptable delays in data collection and analysis, skin prick testing was removed from the protocol. The survey and survey areas were modified to comply with HIPPA. Despite being an expedited protocol, it took several months of working directly with the IRB to gain final approval for the study. At that point, 500 surveys were sent out to addresses randomly sampled from the study areas. When it became apparent that the response rate would not result in a significant sample size, the plan to travel to North Carolina and obtain spirometry measurements was cancelled.
Methods: Two study areas were selected from US census tracts. The exposure area was found in Duplin County, in an area of high CAFO concentration. The control area was located in Hoke and Scotland Counties, which have a relatively low concentration of CAFOs, most of which are poultry, rather than swine, and which produce less air pollution. Addresses were sampled randomly from these census tracts and 500 surveys mailed to 250 addresses in each study area.
Results: Five hundred survey were mailed. Twenty-seven were returned as undelivered. There were seven respondents from Hoke and Scotland Counties. Four reported a physician diagnosis of asthma and three did not. There were five respondents from Duplin and Scotland Counties. One reported a physician diagnosis of asthma and four did not. Of these respondents, none reported hospitalizations, use of prednisone, or poor control of their asthma.
Discussion: The impact of CAFO pollution on neighboring communities remains unclear. The aims of this study were to examine markers of atopy in neighboring communities that might be statistically significant on a small scale. While samples of this nature and in areas with demographics such as in this study are expected to have a poor response rate, the response rate of this study was unusually poor. There were multiple factors believed to have contributed to this. First, the demographics of the study areas were low income, low literacy, high percentage African-American, and rural. Second, this is a highly politicized issue in North Carolina. Third, many features of the survey were changed during the IRB approval process, resulting in a survey that was somewhat more confusing and less readable.
This study provided insight into the difficulties of studying this issue using address based sampling or other survey methods. In the future, further studies with larger budgets may rely on address based sampling and obtain better results. It is likely that as new technologies, such as smart phones and other devices emerge as tools for research, that these will provide better avenues for surveying populations in rural areas with these demographics.
Last updated: 9/22/2014
Updated on September 22, 2014.