Former McHenry County epidemiologist Sherrie Gallas found it intriguing – but not statistically significant – that three next-door neighbors in a town of 1,000 people contracted two rare forms of brain cancer.
Three days after a Philadelphia attorney sued two Ringwood manufacturers, accusing them of causing the neighbors’ cancers through polluted air and water, Gallas e-mailed her supervisors at the McHenry County Department of Health. She concluded in the April 28, 2006, e-mail that neither cancer was as rare as was being reported.
It was the first step in a monthlong health department study that would conclude nothing was out of the ordinary – a conclusion that both the plaintiffs’ attorney and a Northwest Herald investigation are calling into question.
Gallas wrote that the lone case of glioblastoma was not surprising compared with the brain cancer rates for the village’s ZIP code reported by the Illinois State Cancer Registry, because the disease makes up 60 percent of all cancers that start in the brain. She made a similar judgment on the two cases of oligodendroglioma, which makes up about 3 percent of primary brain cancers and occurs in fewer than one person in 100,000.
“The one [glioblastoma] case reported in this case is certainly not all that unusual,” Gallas wrote.
However, the type of glioblastoma that killed plaintiff Franklin Branham is as rare as the news stories and lawsuits were reporting. In her e-mail, Gallas erroneously cited the 60 percent rate for all four stages of glioblastoma, not the much lower rate for Branham’s glioblastoma multiforme. The disease makes up 26 percent of primary brain cancers, or about three cases per 100,000 people. Gallas acknowledged the error in a deposition she gave as part of the McCullom Lake lawsuits.
What’s more, the state data available to Gallas ended in 2003, three years before the lawsuits were filed, and before any of the first three plaintiffs were diagnosed.
After several weeks of investigating the allegations of a cancer cluster, the health department presented its findings to the Board of Health on May 22, 2006, and to worried village residents nine days later at McCullom Lake Village Hall, this time accompanied by the Illinois Environmental Protection Agency. In between, three more brain cancer victims filed suit.
While acknowledging the age of the available data, county officials concluded that local brain cancer rates were not above average. They assured village residents that the companies accused of causing the cancers – Rohm and Haas, its subsidiary Morton International, and Modine Manufacturing Co. – had nothing to do with them. And the county concluded that groundwater contamination from the factories, which the companies had been tracking since the early 1980s, never reached the village’s wells.
Since then, the plaintiff’s attorney has deposed Gallas, the Northwest Herald has interviewed department officials, and obtained department research and communications under the Freedom of Information Act. And the newspaper’s analysis of the data suggests that the county’s investigation and subsequent all-clear was premature, inadequately researched, and incomplete.
It reveals that the cancer cluster study, Gallas’ first, was guided by two of her college textbooks, class notes and Web sites. The analysis relied on cancer data last updated three years prior, before most of the plaintiffs were diagnosed, and the county’s data were limited to McCullom Lake’s ZIP code, not taking into account victims who moved elsewhere.
County officials acknowledged that they didn’t follow the steps that they set down to investigate the cancers, and aside from a round of private well testing and several letters to government agencies, based their conclusions regarding the contamination on the maps from the defendant companies themselves. The county also focused solely on groundwater, disregarding allegations that residents were breathing contaminated air.
In short, Gallas’ conclusions in the April 28, 2006, e-mail were but a few of many mistakes made by the health department, plaintiff attorney Aaron Freiwald said. And he said the resulting study short-changed county residents and taxpayers who deserve better.
“I understand the county’s interest, but what they chose to do in response to a very serious public health concern is go to the companies, meet with them privately, take their documents and diagrams, and show them to the community and say nothing’s wrong,” Freiwald said. “I think that’s wrong. But that’s exactly what they did.”
Plaintiffs who attended Gallas’ deposition said their disappointment in the county’s conclusions had only deepened with knowledge of the county’s methodology.
“It amazed me,” class-action plaintiff Donna Gates said. She laughed. “And not in a good way.”
A year and a half later, county Public Health Administrator Patrick McNulty and Nursing Director Fran Stanwood stand by the county’s findings. Gallas has taken another research job since the study. She, too, stands by her investigation but declined to comment further.
The day before McNulty’s and Stanwood’s Oct. 2 interview with the Northwest Herald, four more plaintiffs filed suit, bringing the number of lawsuits to 22, all but four of them on behalf of brain or nerve cancer victims.
Gallas’ statistical analysis concluded that the 14 brain and nerve cancers reported for the village’s ZIP code from 1999 through 2003, and the number of cases going back to 1989, were not above average.
Postal codes are the smallest unit available to the county from the state cancer registry. McCullom Lake’s population accounts for just over 2 percent of the ZIP code, which includes all or parts of McHenry, Johnsburg, Lakemoor and other cities.
“The data we looked at was conclusive; there’s no doubt about that,” said McNulty, a 29-year department veteran. “It does give us very specific information on the number of cancers that have happened in that location, happened in our county, happened in the state of Illinois. That’s very conclusive information.”
But one reason nothing out of the ordinary was found could be that the county’s data excluded all but one of the current plaintiffs, who either were diagnosed after the state data ended or who lived outside McCullom Lake’s postal code.
To show up in Gallas’ study, a brain cancer victim would have to be diagnosed between 1989 and 2003 while living in the 60050 ZIP code. Of 18 brain and nerve cancer plaintiffs, the only one who meets both criteria is Michael Smulski, who was diagnosed with meningioma in 2002 while living in McHenry.
Six other victims were diagnosed in the 15-year period, but none of them lived in the ZIP code at the time. The remaining 11 were diagnosed after the data set available to Gallas ended, four of them since the county released its analysis.
As of today, five of the plaintiffs with brain cancer, including Smulski, live in the 60050 ZIP code. Four of them live in McCullom Lake village boundaries.
“I think this shows the limitations of the approach the county tried to use, which were obvious at the time and are even more obvious today,” Freiwald said. “We’ve been saying from the beginning that one of the truly striking features of this brain cancer cluster is how many people [were affected] in a very close geographic area and very small time period. You have to look at this for what it is, which is off-the-charts extraordinary.”
While the health department was preparing to conclude that the county’s cancer rates were not above normal, the Illinois Department of Public Health’s chief epidemiologist thought otherwise.
In a May 3, 2006, e-mail written by an IDPH spokeswoman preparing to respond to another newspaper’s questions, epidemiologist Dr. Tiefu Shen concluded that although the county’s glioblastoma rate was below the state average, the oligodendroglioma rate was higher and “statistically significant.”
This e-mail was not forwarded to the county health department, and Gallas testified in her September 2006 deposition that she was never made aware of Dr. Shen’s opinion in her conversations with state health officials.
Missed steps and missteps
One of the most basic tenets of any epidemiology study is to first develop a protocol, according to accepted international conventions on the study of disease. A protocol is a blueprint that not only establishes the experiment’s standards, but also allows the study to be peer reviewed, and results to be replicated.
But it is a step that neither Gallas nor her supervisors took. A protocol was not included in their research obtained by the Northwest Herald, and Gallas testified in her deposition that the county conducted its research without one.
But as far as missed steps go, Gallas’ deposition and subsequent Northwest Herald interviews with county health officials reveal that county investigators skipped many that they, in fact, set down for themselves.
In earlier drafts of the county’s presentation, one slide never seen by the public listed eight steps needed to conduct a cancer cluster investigation. Many of these steps were not finished or not conducted at all, according to both Gallas’ deposition and interviews with county health officials.
For example, the first step mandates describing the cluster – besides never writing a protocol, Gallas never described the cluster in writing during the research, she testified. McNulty said a protocol was not necessary, because the department was merely analyzing existing data. McNulty repeatedly referred to the study as a “statistical analysis” rather than an epidemiology study.
“It wasn’t really a study,” McNulty said in a subsequent interview. “We were using existing information. We weren’t creating new information – it was just what was taken from the Illinois State Cancer Registry.”
The second step stated that the victims must be interviewed. Gallas testified that the department did not do so, citing medical privacy laws.
After the health department presented its report to village residents, it mailed out a community survey to all McCullom Lake households, hoping to bridge the three-year gap in the state’s data, McNulty said. But of the 399 surveys successfully delivered, only 11 were returned, or about 2.8 percent.
“There was really nothing additional brought out in that information that would lead us to change our views at this point in time,” McNulty said.
Gallas testified that she asked for advice on conducting a community survey from both the state cancer registry and a colleague at the University of Illinois medical school. Both advised her not to do so, saying that the results of such surveys are “inconclusive, incomplete, and often cannot be analyzed.”
McNulty said that the department made several unsuccessful attempts to get the plaintiffs’ medical records from Freiwald. Stanwood said they would be an important step in identifying causality.
“There’s more than chemical causes for cancer,” she said. “You are taking into the fact of heredity, where people have lived, what their hobbies are, what they have consumed. There are many different factors.”
Freiwald countered that he received only one request from the county for his clients’ medical records, dated Dec. 15, 2006, months after the study. In Freiwald’s response letter, he declined, pending a request that “would serve a legitimate scientific inquiry.”
Where the wind blows
Steps three and five on the county’s slide mandate assessing the exposure risks. However, the county investigated one exposure route, but did not investigate the other.
Although county officials highlighted their recent well tests and defendants’ maps as proof that village well water had never been tainted, they ignored the lawsuits’ allegation that residents had been exposed to the chemicals through the air. The lawsuits claim that accidents, natural evaporation, and the former air-stripping system installed to clean the groundwater created an inhalation hazard.
Gallas testified that health department officials did not look into allegations that air contamination had reached the village. The allegation is contained in the lawsuits, but Gallas testified she did not read them until the day before the village meeting, after McNulty told her in an e-mail that the suits were available on the Web site of Freiwald’s law firm.
But McNulty said the department discredited the air theory because the factories were north of the village, and the wind in the county blows from west to east.
“I live here,” McNulty said.
“I watch the weather every night, and I see our storms come from the west, and come across McHenry County, and keep moving to the east, and that’s typically how our weather is. Now I don’t have a number to put on that, to say it’s 75 percent of the time, or 80 percent or 60 percent.”
But wind from the north is not particularly uncommon, according to wind reports cited by experts retained by both Freiwald and the defendants.
Five-year measurements of wind speeds at O’Hare International Airport between 2001 and 2005 show that westerly winds are by far the most common, but winds from the north still account for at least 15 percent of the total.
The on-site wind meter at the Rohm and Haas plant in 2005, the year with the most complete data, showed winds from straight north almost as frequent as wind from straight west, although wind from all westerly directions was much more common.
Air pattern experts on both sides factored in more than the manufacturing plants when they charted possible air emissions. Expert reports also factored in natural evaporation from the plume of contaminated shallow groundwater coming from the factories.
Rohm and Haas/Morton installed an air stripper, which pumps out contaminated groundwater and exposes it to atmosphere, because volatile organic compounds readily evaporate, according to U.S. Environmental Protection Agency fact sheets on the chemicals listed in the lawsuits.
But Gallas testified that she never acquired a list of the chemicals involved, or studied their effects on humans, during the research leading up to the village meeting. The sixth step on the county’s slide mandates assessing the likelihood that the chemicals were a factor.
An April 28, 2006, e-mail from Gallas to Environmental Health Director Patti Nomm asking for a list of the chemicals in question was never returned, Gallas testified in her deposition. She also testified that she decided to focus on the statistics and leave the chemicals to other staff.
But staff concluded that the chemicals were not important, Gallas testified and McNulty said in his interview, because the maps from the defendants showed that the contamination plume drifting from both factories never touched village wells.
Gallas further testified that once she concluded that local cancer rates were not above average, she “ ... didn’t feel it was necessary to continue with any additional analysis.”
The department’s final report, dated August 2006, does list the chemicals in question.
Stanwood and McNulty said the slide outlining the eight steps was removed to improve the flow of the village meetings, and denied that it was deleted because the steps listed were not followed. But Freiwald said the county at the very least is guilty of undue haste that he contends was an effort to conclude that nothing was amiss.
“The county, I believe, really jumped the gun on this situation in the name of public health, while really doing nothing in the name of public health,” Freiwald said.
Maps and gaps
McNulty said the chemicals truly are irrelevant because the multitude of research maps from certified experts hired by Rohm and Haas/Morton showed the contaminated groundwater well north of the village.
“This is one of my favorites,” he said, pulling out a map from URS Corp., one of the many research firms retained by the company over two decades. The yellow-hatched plume on the map is nowhere near the town and its domestic wells.
The report from which the map is pulled can be found on the health department’s Web site, which also contains links to other information compiled by the county that conclude that the factories never polluted village wells. None of the expert reports commissioned by Freiwald, which were released in July and are public documents because of the class-action lawsuit, are listed. And Freiwald said he has never been asked by the county for any of them.
McNulty said the county never saw a problem with relying on maps from the defendants.
“To say that the groundwater contamination from Modine and Morton and Rohm and Haas caused some type of illness outbreak in McCullom Lake, I would say would be completely inaccurate, and that was one of the allegations that had been made to that point,” McNulty said.
The county tested nine village wells after the lawsuits were filed, and only one tested positive for a volatile organic compound, which was not listed in the lawsuits and well below federal safety standards, according to an October 2006 letter from the IDPH. The letter concludes that the contaminants have not reached the village, but states that the conclusion was reached based on the defendants’ maps.
As for groundwater flow, the county contacted the Illinois State Water Survey, the agency that mapped McHenry County’s aquifers in 1998. In a Sept. 8, 2006, letter, months after the county presented its conclusions to residents, senior geologist H. Allen Wehrmann likewise said that it would not be possible for contaminated groundwater from the factories to flow to the village. However, Wehrmann’s conclusion on the location and extent of the contamination is based on the defendants’ maps.
Freiwald’s experts contend that contamination reached a deep aquifer that transported it to village wells, which the defendants dispute.
Too close for comfort
Freiwald attended the May 31, 2006, village meeting in which the health department presented its findings. He said more than a few audience members noted who provided some of the color maps and diagrams that the county used to make its case.
“There were people in the room who were gasping when they saw them, because in the corner, [the maps] said courtesy of Rohm and Haas,” Freiwald said.
And while Rohm and Haas shared its maps with the county health department, the county shared its report with company officials before showing it to village residents.
Rohm and Haas Ringwood facilities manager Tom Bielas testified in his October 2006 deposition that health officials showed him and corporate spokesman Syd Havely the final presentation before residents saw it. Bielas did not recall company representatives recommending any changes.
Both McNulty and Havely said that company officials were just dropping by, and the county offered to show it to them. Freiwald said he was not so sure that the visit was happenstance, given that he did not get a similar invitation. At the very least, Freiwald said, going over the study with one of the defendant companies before the public saw it calls its integrity into question.
“They didn’t meet with me. They didn’t contact me and say: ‘Hey, we met with Rohm and Haas, we’d also like to meet with you. We heard their side of the story – we’d like to know, maybe, you have some information or some documents or some data that will round out the information that we have,’ ” Freiwald said.
In the 20 months since the lawsuits, for example, a memo has surfaced proving that Morton knew about the contamination a decade before the company reported it. Also, a December 2006 report indicates that the deep aquifer contains contaminants that can be linked to the landfill.
“They took Rohm and Haas’ data; they took Rohm and Haas’ word for it,” Freiwald said. “Period. End of story.”
McNulty said that he did not think that sharing the county’s findings with Rohm and Haas officials before showing them to the public tainted the study’s credibility. He also said that Freiwald was not contacted because the health department decided that the information it had was conclusive.
“I didn’t see there was any need to do that, because our information indicated the flow of groundwater in McHenry County,” McNulty said.
Stanwood, the county nursing director, steadfastly denied that any corporate or government pressure influenced the department’s findings.
“Absolutely not,” she said.
But Freiwald claims that, from start to finish, the county’s study was a fatally flawed attempt to absolve the companies of blame.
The first word that came to the mind of plaintiff Scott Milliman, a county sheriff’s deputy diagnosed in 2002 with oligodendroglioma, is not suitable for print.
“What a joke,” said Milliman, who attended Gallas’ deposition.
“Unprofessional” was the first word that came to the mind of plaintiff Brian DiBlasi, who also attended the deposition. He said that not enough was done, and in too quickly of a time, for the study to have any merit.
The fact that three next-door neighbors contracted brain cancer is all the proof that plaintiff Robert Nelson said he needed. He said he thought that county officials jumped to conclusions too soon, especially since three more lawsuits were filed the week before the presentation to the village.
“What are the odds of people getting these unpronounceable brain tumors living in such a small area? And for the county health department to come out – it certainly seems premature – to sound an all-clear and say there’s no problem?” Nelson said. “That certainly seems surprising to me. They’re supposed to be acting in the public interest.”
County officials insisted that the public interest was what they had in mind all along.
McNulty said a follow-up study was not out of the question once more updated data came from the state cancer registry, which has data through 2004 as of today. Four of the brain cancer plaintiffs were diagnosed in 2004, but only one – Bryan Freund – lived in the 60050 ZIP code at the time.
“Sometime in the near future, we’ll have additional years present that will need to be looked at, but the trend to this point would indicate [that] our conclusions are consistent with what the data supported at this time,” McNulty said.
Freiwald suggested that the health department should sit on the sidelines while the legal process slowly moves forward because it does not have the expertise needed to draw a conclusion one way or the other.
“It’s what I say to every client before a deposition – if you don’t know, say you don’t know,” Freiwald said. “Don’t make up something, because it just hurts your credibility, and I think that’s what happened here.”