Cancer ‘hot spots’ may be linked to hazardous waste sites

Researchers identified a potential association between pockets of increased cancer incidence among adults in Florida and their proximity to hazardous waste sites.

Emily Leary, PhD, assistant professor at University of Missouri School of Medicine, and Alexander Kirpich, PhD, postdoctoral associate at University of Florida, examined rates of all cancer types among adults reported by the Florida Department of Health from 1986 to 2010. The study excluded pediatric cancers, as they are likely to have other nonenvironmental causes.

Leary and Kirpich attempted to identify clusters of cancer incidence — or “hot spots” — and assessed whether presence of those hot spots correlated with locations of hazardous waste sites, known as Superfund sites.

The results — published online in Statistics and Public Policy — showed cancer incidence rose by more than 6% (95% CI, 3-9) in counties that had Superfund sites.

HemOnc Today spoke with Leary about the findings and their potential implications.

Question: What prompted you and your colleagues to conduct this study?

Answer: It came about from what we had seen and heard in the news locally. My colleague and I were both at University of Florida, and there is a Superfund site in Gainesville. There were several local news articles about the sample testing that had occurred at this site, so we got curious. We had never really thought about what health issues might arise due to these sites, but we bounced some ideas off a spatial epidemiologist we were working with and it grew into this study.

Q: How did your study differ from others that examined potential association s between cancer and proximity to hazardous waste sites?

A : Much of the research done in the past has looked at specific types of cancer, and they have been smaller-scale, shorter time-frame studies. This study examined all cancer types over an extended period. Many earlier studies didn’t have access to a longitudinal set of data that encompassed as many years as ours did. Also, many of the geographic information systems and spatial modeling techniques have developed in ways that weren’t available when many other studies were done.

Q: How did you conduct the study?

A: We gathered information on Superfund site locations from the U.S. Environmental Protection Agency’s ToxMaps website. That site contains public data that provides the name, location and toxicology indicators for each Superfund site. We then asked the Florida Cancer Data System for adult cancer incidence data from 1986 to 2010. One of the major challenges was that we were working with data from a single location — also known as point-level data — from the Superfund sites, and the cancer incidence data were provided on a county level. We had to reconcile those spatial measures and aggregate the data to one spatial level. Further, most pollution studies are done on air pollution, but we wanted to focus on a different method of pollution transport — water. This was especially important in Florida, because the state has very shallow water aquifers. Once we aggregated the data, we looked at surface water location in relation to each Superfund site and came up with a measure of how accessible water was with respect to each of these Superfund locations.

Q: Can you describe the key findings?

A: Our research identified these cancer ‘hot spots’ and located them with respect to both men and women. One surprise was that a lot of the clusters aren’t exactly at the location of the Superfund sites. This could be due to several factors. It could be because we had to aggregate the data at multiple spatial levels. We had data from multiple sources and we had to combine them in some way, which introduces error. Also, when you study cancer — particularly among adults, for whom cancer is generally thought to have a long latency period — it is difficult to accurately assign an exposure. You are trying to follow them for decades but you really can’t get a precise residential history and other data that may be needed. There is a lot of ecological bias, and during the period we studied, there was a huge population influx in Florida. Some of these strange findings that we originally did not expect likely are due to these outside influences on the data. It is very difficult to account for these influences. In our case, there was a huge population influx into Florida from 1986 to 2010. Beyond accounting for population differences during this time, additional information that is not available or recorded would be needed.

Q: How can your results inform the public, clinical community or public health agencies?

A: That's the big question: How does this make a difference? This type of research can help agencies and policymakers focus on specific areas that may need more investigation or more resources. It is a way to narrow their focus a bit and make them more efficient in funding and spending, and provide focus on areas that may warrant further investigation.

Q: Is there any additional research that needs to be conducted or key questions that need to be answered before definitive conclusions can be drawn?

A: This is a very high-level study. We examined the data at a very coarse level. The real value, now that we know there is something here that warrants further investigation, is to drill down a bit and figure out what exactly that ‘something’ is. The first step probably would be to look at this problem in terms of specific cancer types and not all of them together. – by Kyle Doherty

Reference:

Kirpich A and Leary E. Statistics and Public Policy. 2016;doi:10.1080/2330443X.2016.1267599.

For more information:

Emily Leary, PhD, can be reached at learye@health.missouri.edu.

Disclosure: Leary reports no relevant financial disclosures.

Researchers identified a potential association between pockets of increased cancer incidence among adults in Florida and their proximity to hazardous waste sites.

Emily Leary, PhD, assistant professor at University of Missouri School of Medicine, and Alexander Kirpich, PhD, postdoctoral associate at University of Florida, examined rates of all cancer types among adults reported by the Florida Department of Health from 1986 to 2010. The study excluded pediatric cancers, as they are likely to have other nonenvironmental causes.

Leary and Kirpich attempted to identify clusters of cancer incidence — or “hot spots” — and assessed whether presence of those hot spots correlated with locations of hazardous waste sites, known as Superfund sites.

The results — published online in Statistics and Public Policy — showed cancer incidence rose by more than 6% (95% CI, 3-9) in counties that had Superfund sites.

HemOnc Today spoke with Leary about the findings and their potential implications.

Question: What prompted you and your colleagues to conduct this study?

Answer: It came about from what we had seen and heard in the news locally. My colleague and I were both at University of Florida, and there is a Superfund site in Gainesville. There were several local news articles about the sample testing that had occurred at this site, so we got curious. We had never really thought about what health issues might arise due to these sites, but we bounced some ideas off a spatial epidemiologist we were working with and it grew into this study.

Q: How did your study differ from others that examined potential association s between cancer and proximity to hazardous waste sites?

A : Much of the research done in the past has looked at specific types of cancer, and they have been smaller-scale, shorter time-frame studies. This study examined all cancer types over an extended period. Many earlier studies didn’t have access to a longitudinal set of data that encompassed as many years as ours did. Also, many of the geographic information systems and spatial modeling techniques have developed in ways that weren’t available when many other studies were done.

Q: How did you conduct the study?

A: We gathered information on Superfund site locations from the U.S. Environmental Protection Agency’s ToxMaps website. That site contains public data that provides the name, location and toxicology indicators for each Superfund site. We then asked the Florida Cancer Data System for adult cancer incidence data from 1986 to 2010. One of the major challenges was that we were working with data from a single location — also known as point-level data — from the Superfund sites, and the cancer incidence data were provided on a county level. We had to reconcile those spatial measures and aggregate the data to one spatial level. Further, most pollution studies are done on air pollution, but we wanted to focus on a different method of pollution transport — water. This was especially important in Florida, because the state has very shallow water aquifers. Once we aggregated the data, we looked at surface water location in relation to each Superfund site and came up with a measure of how accessible water was with respect to each of these Superfund locations.

PAGE BREAK

Q: Can you describe the key findings?

A: Our research identified these cancer ‘hot spots’ and located them with respect to both men and women. One surprise was that a lot of the clusters aren’t exactly at the location of the Superfund sites. This could be due to several factors. It could be because we had to aggregate the data at multiple spatial levels. We had data from multiple sources and we had to combine them in some way, which introduces error. Also, when you study cancer — particularly among adults, for whom cancer is generally thought to have a long latency period — it is difficult to accurately assign an exposure. You are trying to follow them for decades but you really can’t get a precise residential history and other data that may be needed. There is a lot of ecological bias, and during the period we studied, there was a huge population influx in Florida. Some of these strange findings that we originally did not expect likely are due to these outside influences on the data. It is very difficult to account for these influences. In our case, there was a huge population influx into Florida from 1986 to 2010. Beyond accounting for population differences during this time, additional information that is not available or recorded would be needed.

Q: How can your results inform the public, clinical community or public health agencies?

A: That's the big question: How does this make a difference? This type of research can help agencies and policymakers focus on specific areas that may need more investigation or more resources. It is a way to narrow their focus a bit and make them more efficient in funding and spending, and provide focus on areas that may warrant further investigation.

Q: Is there any additional research that needs to be conducted or key questions that need to be answered before definitive conclusions can be drawn?

A: This is a very high-level study. We examined the data at a very coarse level. The real value, now that we know there is something here that warrants further investigation, is to drill down a bit and figure out what exactly that ‘something’ is. The first step probably would be to look at this problem in terms of specific cancer types and not all of them together. – by Kyle Doherty

PAGE BREAK

Reference:

Kirpich A and Leary E. Statistics and Public Policy. 2016;doi:10.1080/2330443X.2016.1267599.

For more information:

Emily Leary, PhD, can be reached at learye@health.missouri.edu.

Disclosure: Leary reports no relevant financial disclosures.