By now many Americans have heard of the once-obscure Boston suburb of Woburn, thanks to Jonathan Harr’s bestselling 1995 book, A Civil Action, and the 1998 movie of the same name starring John Travolta. Both describe the civil suit brought against Beatrice Foods Company and W.R. Grace & Company by Woburn families whose children suffered from leukemia. The corporations were accused of contaminating the two wells that provided much of the water supply to eastern Woburn, where most of the families lived.

The movie depicted the rise and fall of Jan Schlichtmann, an upstart personal injury lawyer with a penchant for exorbitant spending, who was hired by the families to battle the two huge corporations and Boston’s legal establishment. But it missed the scientific story line that is an integral component of Harr’s book. Two faculty members in the School’s Department of Biostatistics, former chair Marvin Zelen and incoming chair Stephen Lagakos, played leading roles in that drama; they were the first to find statistical evidence linking Woburn’s elevated childhood leukemia rates to the contaminated wells.

When Zelen, Lagakos, and a research assistant began their study of Woburn, they knew the difficulties they would face in trying to identify a possible cause of elevated cancer rates in a small community. Cancer is caused by multiple factors and takes years to develop (though longer in adults than in children). Additionally, in many such "cancer clusters," different kinds of cancer with entirely different causes are often lumped together. The Woburn children all had leukemia, but they did not all have the same type of leukemia. Even now, the risk factors for leukemia are not well understood.

And then there were the numbers. Twenty cases of childhood leukemia were identified in Woburn between 1964, when the two wells servicing eastern Woburn (labeled wells G and H) began pumping, and 1983, the time of Zelen and Lagakos’s study. Such a small number makes identifying a statistically significant link between the cancer cases and any particular factor very difficult. In studies of tobacco use and many occupational hazards, researchers acquire larger sample sizes by extending their studies across communities. But as Lagakos points out, the mix of contaminants in wells G and H "was unique to Woburn." The researchers knew that a study of Woburn’s excess leukemia couldn’t be replicated on a larger scale.

But what if leukemia wasn’t the only harmful health effect potentially related to the well water? Zelen and Lagakos conjectured that "if there really is something bad going on, affecting children and inducing leukemia, it must also cause other diseases," says Zelen. "We thought about it and decided that the most vulnerable population was the fetus." So in addition to studying the relationship between access to G and H water and childhood leukemia incidence, they studied the incidence of adverse pregnancy outcomes--spontaneous abortions, low birthweight--and 14 categories of childhood disorders. "Up until that time no one had suggested that congenital abnormalities and birth defects would be related to the [Woburn] well water," Zelen says. Data for the study was collected by over 300 volunteers, including Woburn citizens and members of the School of Public Health community, who conducted a phone survey of more than 5,000 households in the city of 37,000 people.

Using a report by the Massachusetts Department of Environmental Quality and Engineering that detailed the water distribution pattern of Woburn’s eight wells and estimated how much water each of five Woburn regions received from G and H, Zelen and Lagakos deduced what annual percentage of a household’s water had come from the two contaminated wells. Wells G and H had most likely never serviced west Woburn, where about two-thirds of the city’s population lived. In east Woburn, the amount of G and H water varied by time and place.

The calculations of exposure to contaminated well water were less than precise. Even if a household received most of its water from G or H or a combination of the two, "we don’t know whether people drank water from the wells or whether they had bottled water," or even if they drank most of their water at home, or how long the water was contaminated, Zelen says. But by estimating the amount of contaminated water individuals might have received and comparing the exposed to the unexposed, the researchers came to a conclusion that was probably both reassuring and disturbing to the people of Woburn. In February 1984, they announced their highly anticipated findings at a community meeting held at the Woburn Trinity Episcopal Church.

"We found that the children who had leukemia…had access to twice as much water as the children without leukemia," Zelen says, or, put another way, leukemias "tended to congregate" in households with more water from G and H, as opposed to households that got little or no water from the two wells. The study also found positive associations between access to G and H water and both perinatal deaths and a category of adverse birth outcomes that included central nervous system, chromosomal, and oral cleft anomalies, as well as positive associations with childhood disorders involving the urinary tract and the respiratory system.

To see if shutting down the wells had had any effect on health in Woburn, Zelen and Lagakos compared the rates of birth defects before and after the wells’ shut-down in 1979. "And sure enough," Zelen remarks, "you could see that after the wells were turned off, there were no or very few birth defects." The same was true before 1979 when comparing east Woburn’s birth defect rates during times when the wells were pumping to rates at times when the wells were shut off. The latter finding, Zelen argues, counters one of the criticisms leveled at the study: that some uncontrolled-for factor specific to east Woburn might have accounted for its elevated childhood leukemia rates.

So, you might ask, were the wells the real culprits? Both Zelen and Lagakos repeatedly emphasize that cause and effect cannot be proven by a study like theirs. Zelen says, "Our findings were consistent with an association which resembled cause and effect." That isn’t a completely satisfying answer, but Lagakos cautions that finding an association between disease and cause is, "in some respects…as much as science can do" under these circumstances. They will say that the greater the amount of well water a household received in any given year, the higher the likelihood that leukemia or certain birth outcomes or childhood disorders would appear in that household. "Does it prove it was the wells?" asks Lagakos, answering quickly, "Absolutely not."

The study could say nothing about the specific contaminants in the well water that might have been responsible for the illness in Woburn or the source of those contaminants. Those questions were tackled by others. In 1986 the case against Beatrice was dismissed, and Grace made an $8 million settlement with the Woburn families. Later, an EPA report concluded that both companies had contaminated the city wells. Grace has engaged in an aggressive PR campaign to highlight its efforts to clean up its property in Woburn. The Massachusetts Department of Public Health released two studies on Woburn in recent years, the Woburn Environment and Birth Study (WEBS) in 1996 and the Woburn Childhood Leukemia Follow-Up Study in 1997. WEBS did not support an association between the contaminated water and adverse pregnancy outcomes and birth defects, while the leukemia study (which followed up a 1981 MDPH case-control study) corroborated Lagakos and Zelen’s findings.

But the most important postscript in Lagakos’s mind? Cancer rates in Woburn are no longer elevated.