Study focuses on clusters to find cause of Del.’s high rates
By JEFF MONTGOMERY
The News Journal
A quarter century after Delaware found itself with the nation’s highest cancer death rate, health officials have narrowed their search for cancer clusters to the neighborhood level, where for the first time local rates of rare cancers could provide clues to environmental or other causes.
In one of the most unique studies of its kind, state epidemiologists identified 45 out of 196 zones with total cancer rates exceeding state and national rates by margins health officials regard as significant.
Some of those cancers are of a type known to have environmental links, an ominous discovery in a state with a national legacy of asthma-inducing air pollution, toxic waste dumps and groundwater contamination that has now breached an aquifer serving tens of thousands in Delaware, New Jersey and Maryland.
Across the state, cancers most often found to be significantly elevated were breast, colorectal, lung and prostate, already the most common types in Delaware and nationwide.
But in some spots, rarer cancers emerged at rates far higher than state averages, even after the Division of Public Health made large allowances for margins of error due to age and small sample sizes in U.S. census tracts with an average of 4,000 residents. Researchers examined reports of all cancers from the most recent years available, 2002-2006.
Among the worst neighborhood-level clusters, they found:
• Laryngeal cancer cases in Northeast Wilmington next to Brandywine Creek between Northeast Boulevard and Market Street.
• Melanoma in a small tract around Brandywood, south of Naamans Road between Grubb and Foulk roads.
• Ovarian cancer rates in the Hockessin area that were more than four times the state average.
• Thyroid cancers in the Overview Gardens and Minquadale areas southwest of Memorial Drive.
• Leukemia in a suburban and rural tract west of Dover and northwest of Wyoming.
• Ovarian cancer east of Bridgeville and Greenwood.
• Melanoma, kidney and esophageal cancers west of Rehoboth Beach.
Census tracts with unusually high cancer rates turned up in all three counties, from the top of Brandywine Hundred to west Dover and the Love Creek area in coastal Sussex County.
Epidemiologists looked for four patterns: the rates of all types of cancers in a single tract, the rates for specific cancers in those tracts, differences between men and women, and average ages of diagnosis, especially early or late.
The top 10 average rates for cancers of all types and the highest rates for men and women were confined to New Castle County and Kent County, but scattered tracts in all three counties had unusual spikes of specific cancers.
The tract-level cancer study was ordered by the Delaware General Assembly after The News Journal waged a long struggle to persuade state health officials to release detailed cancer rate data. Delaware now is one of the few states in the nation to assess cancer rates for every census tract. To prevent identifying any one person with cancer, health officials released only average annual rates for all cancers and did not list the actual counts, which range from one or two new cases a year to a few dozen.
The Centers for Disease Control has investigated and confirmed dozens of clusters in dozens of states over the years. But federally led investigations seldom turn up a single, clear-cut cause. Cancers can have many, hard-to-verify triggers, the agency pointed out, including family background, lifestyle choices such as smoking and eating habits, occupational exposures and pollution encountered in communities far removed from the home of record where the cancer was diagnosed.
How far Delaware researchers will continue to explore the data depends on money and political will.
“There are a lot of red flags here,” said Dr. Paul Silverman, Public Health’s associate deputy director for health information and science. “I would agree with the concern. However, as is true with all of these census tracts, all you can do at this point is hypothesize. We would actually have to have additional information in order to answer any questions.”
Department of Natural Resources and Environmental Control Secretary Collin P. O’Mara said his agency is reviewing the results and will take the small-area cancer statistics into account while setting priorities in the future.
A lack of evidence of any direct tie to air or water toxins will not inhibit his agency from taking steps to reduce pollution, he said.
Environmental health concerns were behind DNREC’s recent push to curb releases from larger pollution sources, including NRG’s Indian River power plant, O’Mara said. The census tract findings could help focus its efforts on smaller sites or environmental programs that may have a large impact.
Initiatives have expanded
State lawmakers voted in 2008 to require an analysis of each year’s cancer reports by census tract. It was the latest in a series of cancer control initiatives that date as far back as 1986, when then-Gov. Mike Castle announced a push that included fielding mobile mammography vans and talk of statewide environmental reviews.
Castle’s move followed release of grim national health rankings that pegged Delaware as having the nation’s highest death rate from some cancers.
Cancer control services and initiatives have expanded in the years since, and cancer mortality rates have fallen. Yet rates for both deaths and new cancers in Delaware remain higher than the average for the nation as a whole.
In this latest study, overall rates in eight tracts appeared to reach “significant” levels, according to Public Health, because rates for several individual cancers in that area randomly edged higher than the state average.
For the remaining 37 tracts where Silverman said there could be concern, however, rates were significantly high for at least one of 23 types of cancer, including several diseases sometimes tied to environmental exposures, such as bladder, thyroid, liver and brain cancers, leukemia and non-Hodgkins lymphoma.
Bladder cancer can be caused by workplace exposure to chemicals. Liver cancer has been tied to arsenic in water, steroids and on-the-job pollution exposures. Radiation exposure can cause thyroid and brain cancer, as well as leukemia and non-Hodgkins lymphoma, although certain chemicals also can cause leukemia.
Retiree Irene Butcher rarely hears neighbors talk about the cancers that turn up far too often among residents in her Jefferson Farms community, just north of New Castle.
“I haven’t heard any talk, but I read the paper a lot and I do know that Delaware is kind of high up on the list,” for cancer rates, Butcher said. “It seemed like it used to be that if somebody in your family didn’t have it, you didn’t have to worry about it. Now cancer strikes anybody.”
Rates in Jefferson Farms were 70 percent higher than the Delaware average for all cancers — higher than anywhere else — according to Public Health’s study.
Just a few blocks from Butcher, Irene Jones has been battling breast cancer since 2004. Jones described cancer as a personal and family burden, recalling grandparents stricken with or lost to the disease. A retired Colonial School District teacher, Jones said her family originally was from New Jersey, but she also worried that cancer might fall more heavily on the neighborhood where she’s now settled.
“It’s something you’re always concerned about, maybe because of your family history, or where you work, or where you live,” said Jones, who said she has been cancer-free since treatment. “I worry here sometimes because of all the landfills around.”
“People need to be kept informed about cancer, and they needed to be reminded about taking care of their health care,” said Jones, whose cancer was detected by a routine mammogram. “You need to monitor your own health and you need to go for checkups.”
Both men and women had higher-than-average rates for cancer in Jones’ community, prostate cancer was detected at a rate 85 percent higher than the state as a whole, and breast cancer rates were just slightly over the state average.
Push for information
One tract that also stood out in the Public Health study is a swath of land between I-95 and Del. 273, northwest of the Delaware City petrochemical complex.
The zone from Salem Woods in the west to Raintree Village in the east had a cancer rate for men more than double the state average, with an average age at diagnosis eight years younger than was typical.
Prostate cancer rates were 1.7 times higher than the state average. Breast cancer rates were 84 percent higher, laryngeal cancer rates were four times the state’s average and leukemia rates 2.5 times higher.
Leukemia is a disease that state officials noted can have environmental ties, such as exposure to benzene, a common industrial chemical and motor fuel ingredient found in cigarette smoke and auto exhaust. In a yearlong investigation published this year by The News Journal, benzene was one of a number of chemicals found in groundwater across the state.
Salem Woods resident William K. Dadson said the state should press on and quickly share what it finds about the cancer clusters.
“I think it’s good for them to pinpoint it, and let people know what’s going on,” Dadson said. “Cancer is a concern for everyone. It’s very important, and health information should be available to the public.”
State’s rate 9.5% higher
Delaware has attempted several close-up looks at cancer in recent years, an effort partly rooted in nagging, decades-long concerns about the state’s chronically higher average rates of cancer detections and cancer deaths compared with national findings.
State officials struggled through the late 1980s to fend off connections between the state’s cancer rates and its industrial past, high pollution levels and proximity to a so-called “Cancer Alley” in neighboring New Jersey, where high cancer rates led to state and federal investigations.
The state’s most recent five-year report, released in May, found that cancer turned up 9.5 percent more often in Delaware residents than the nation as a whole, with cancer 4 percent more likely to kill.
Despite those sobering numbers, officials have pointed out, the state’s cancer mortality rate has fallen 18.9 percent from a decade ago, when it was among the nation’s highest. The annual rate of new cancer cases has dropped 3.8 percent over the same period.
“Looking at one point in time or one five-year period is not nearly as helpful as looking at trends,” said Dr. Karyl Rattay, Delaware’s public health director.
State officials were prodded again to dig deeper into cancer rates in 2006 and 2007, amid public demands for studies in southeastern Sussex County. Health concerns in that area had become entangled with politically charged debates over offshore wind power, clean energy and emissions from the oil- and coal-burning Indian River power plant near Millsboro.
The result was a widely reported cancer-cluster reflecting higher-than-average lung cancer rates in a large swath of land southeast of the power plant.
Those findings however, were based on ZIP code boundaries, geographic areas that sprawl across several census tracts, some far removed from the power plant.
But it also got the attention of lawmakers troubled by the state’s resistance to releasing cancer rate data by census tract to The News Journal.
State objections to data releases included claims that statistics could be used to identify patients. They also said that such small areas are more likely to be affected by random spikes or drops in cancer cases, distorting true rates.
The Legislature brushed those concerns aside, giving health officials 90 days to produce census tract studies after each year’s statewide cancer report.
Now that the first census tract report has been compiled and broken down further by individual cancers, Heather Woods resident Patsy A. Howaniec says the state needs to share what it knows.
Her neighborhood stands on the edge of a tract newly identified as having the state’s second-highest cancer rate.
“Scary stuff,” Howaniec said. “I’d like to know more.
Lung cancer rates in the tract, north of Du Pont Highway in Bear and including Brookmont Farms and Wellington Woods, were double the state’s average, and oral cancer rates 4.7 times higher.
Small areas produce quirks
Recent findings in Sussex show just how complicated cancer investigations can get in smaller areas, and some of the troubling quirks that they can turn up.
In the new census tract report, neither the greater Millsboro area nor any of the census tracts adjacent to it had unusually high average rates for any cancers, despite lung cancer statistics by ZIP code that inflamed earlier debates over the Indian River power plant emissions.
In fact, Sussex County had the lowest number of tracts with significantly elevated cancer statistics of any in Delaware. In two of the six tracts, no single cancer could be found with notably higher-than-average rates.
Silverman cautioned that Public Health would need more data, time and research to determine whether environmental cancers signal a problem.
“I regard all of these [findings] as clues, which may or may not bear fruit,” Silverman said.
In the case of the suspected lung cancer clusters in the Millsboro area, a series of individual surveys found that higher prevalence of the disease was mainly due to higher prevalence of tobacco use and past work in a range of “high-risk” jobs.
The high-risk job category takes in a huge share of the county work force, including those in jobs involving agriculture, chemicals, construction, manufacturing and pharmaceuticals.
In some areas, clusters represent real puzzles.
Among the tracts with the top 10 highest rates are three areas of Kent County where land use runs the gamut from suburban to small-town and rural — including parts of the county’s small Amish community.
Two tracts in Kent accounted for the state’s fourth- and fifth-highest average rates of all cancers. In some tracts in Kent, cancer was discovered at earlier ages, with high rates for scarce cancers including leukemia and non-Hodgkins lymphoma, which has been linked to radiation, weakened immune systems and some infections.
Some scientific studies also have found correlations between non-Hodgkins lymphoma and exposures to pesticides and emissions from older types of waste incinerators.
Inaccurate population data
Census tract studies are especially tricky, said Elizabeth Ward, the American Cancer Society’s vice president for surveillance and research.
High rates deserve a second look, she said, but a high prostate cancer rate may be evidence of better diagnosis rather than a higher incidence of the disease.
“If you see higher rates of prostate cancer, it may actually reflect higher utilization of screening, said Ward, who lives in Atlanta. “There are some cancers where it’s pretty clear that one of the things really driving the rates is utilization of testing.”
Another complication is the possible inaccuracy of population estimates. Delaware’s census tract populations estimates are based on the 2000 census. The new population count will not be available to the public until February.
Precision especially suffers, Ward said, if several years have passed since a full census tally.
“The smaller the area, the more the chances are that you can get statistical errors,” Ward said. “We saw that toward the end of the time before the last census. We started seeing some strange patterns in cancer rate by county, especially when we looked at cancer rates by county and race.”
In Delaware, Public Health has investigated cancer clusters in small areas of the state for years, based on cancer surveillance reports, requests from residents or other triggers. Studies have never confirmed a cause for any cluster, but some have raised eyebrows.
In fact, an individual resident’s concern about cancer in the Brandywood census tract last year led to a conclusion that melanoma rates among women were significantly higher than the state rate for 2001-2005. That same cluster also turned up in the latest study of the average for 2002-2006, with melanoma more than 330 percent higher than state rates.
A different citizen request this year led to a study of childhood cancer rates and deaths, and found that Kent County childhood cancer rates exceed national rates by a significant amount. In that case, Public Health said it would continue monitoring and “assume position of watchful waiting for childhood cancer incidence trends in Kent County.”
In Salem Woods — south of Old Baltimore Pike near Newark, part of a census tract with the third-highest rate for all cancers — longtime postal worker Earl C. Cunningham has plenty of questions about cancer. Some focus more on how he’s lived than where.
Cunningham was treated for prostate cancer in 1999 but has long wondered whether his sometimes-heavy exposure to the herbicide Agent Orange decades ago during two combat tours in Vietnam set the stage for discovery of the disease. Public Health does not include prostate cancer as having a confirmed environmental cause.
“If there’s a problem in this area, they need to own up to it. But it’s hard to say,” Cunningham said. “Most of the people in this community come and go; they move in and out. The longer-term ones, I’m not aware of them having any cancer-related illnesses.”
Gerard Rushton, a University of Iowa professor who developed Iowa’s cancer mapping program, praised Delaware’s effort but echoed Ward’s concern that random swings can get magnified when the focus gets too tight and the census records stale.
“Once you get down to saying that in this small area we have four or five cancers, the rates go up and down like a yo-yo. You can be looking at a map, but it’s really not a valid map. We call it the small number problem,” said Rushton, who said he is wary of any studies based on zones with fewer than 50 cases.
Rushton, who is awaiting a decision on a National Cancer Institute grant for research on cancer tracking, said that New York and California have developed census tract surveys in recent years. But the practice remains rare in statewide use.
“On the whole, it’s very useful and important to look at data for as fine a geographic area as possible,” said Ward, of the American Cancer Society. “Very often, the factors that determine whether or not there’s a high risk for a specific disease do vary at levels smaller than the county. You may see patterns that you wouldn’t see if you were just able to look at the county level.
“On the flip side,” Ward added, “for some diseases, there are just so few cancer cases that you would be concerned about looking at them at the census tract level, because of confidentiality reasons” and statistical swings.
Proceeding cautiously
Delaware is far from alone in its struggle to use state and national cancer data to improve prevention and detection, Ward said. While census tract records have been available to health agencies for years, public studies using those records are rare.
“I don’t think it’s common. I’m aware that New York did it, but there’s a fine balance with concerns about confidentiality when you bring the data down to that level,” Ward said. She added that high rates can be deceiving.
“It’s very tricky for numerous reasons, not just statistical variations,” Ward said
Rattay said that Public Health is moving cautiously.
Plans are now in the works to brief members of the state Cancer Consortium’s environmental committee, with a blueprint for follow-up action due later this month.
Steps are likely to include talks with community groups and the Department of Natural Resources and Environmental Control on possible reasons for the higher rates, including access to health care and possible sources of pollution exposure.
“We’re going to reach out to the communities in all 45 census tracts and let them know that this report is available, and that we would like to work with these communities to educate them on prevention efforts, screening services and to learn from them about their concerns,” Rattay said. “What this report does is help us to think about where the geographic areas are where we can prioritize our efforts.”
The results also could help the state prepare for more detailed studies of chemical “body burdens,” an idea that also arose out of the Millsboro cancer controversy and past debates over the state’s high rate. Delaware has made several attempts to launch such a study over the years and applied without success for federal help. A body burden study involves an analysis of selected chemical contaminants inside the bodies of residents in a selected part of the state. State officials have estimated in the past that a targeted study would cost as much as $5 million.
“It certainly would be helpful to get that funding,” Rattay said. “That would help us not so much to draw conclusions about whether people are getting cancer from certain exposures, but it would help us identify whether or not certain chemicals were higher in certain individuals’ bodies.”
Jay B. Hammond of Jefferson Farms, whose father died of prostate cancer, said he was unaware the disease was so common around the community.
“I don’t know of anyone here with it, but that’s one of the things that guys don’t talk about,” Hammond said. “With my family history, I keep an eye on it, but I also don’t know why this development would be any different.”
Environmental studies might leave plenty of questions unanswered, Hammond said.
“I don’t know why this development would be any different from one across the street or across the tracks. It’s not like it was built on a toxic waste dump,” Hammond said.
“We’re right up the street from the Delaware City Refinery, down the road from Marcus Hook [refinery] and across the river from a big chemical plant,” he said. “We’re east of everything that goes on in the western part of the country. So what the hell are you going to do?”
The News Journal