Cancer coincidence?

As months went by, Taylor said, she couldn't go two weeks without hearing that another neighbor was sick. She and her neighbors began wondering if a nearby industrial plant could be to blame: It produced fumes and hazardous waste, and perhaps those were causing the disease.

Now Taylor's neighborhood is one of 44 potential "disease clusters" the state Department of Public Health is investigating. But if history is any indicator, she and her neighbors probably never will be able to draw a definite connection between their soil, air or water and the diseases on their street.

Galvanized by widespread public belief that environmental poisons are clearly linked to disease -- and encouraged by a few high-profile cases like those featured in the movies "A Civil Action" and "Erin Brockovich" -- Massachusetts residents have been calling the state at a rate of more than 1,300 a year, worried that something in their environment is causing high numbers of cancers and other diseases.

But an overwhelmed state office can only intensely investigate a few of these calls. And because of deep scientific uncertainty that surrounds their mission, they only rarely -- some say never -- reach a firm conclusion. In fact, some scientists question whether any neighborhood disease cluster has ever been linked to the environment -- even in Woburn, home of the "Civil Action" case, where eight families stricken by leukemia reached an $8 million settlement with a nearby chemical company.

"It is difficult to show that disease clusters are due to an environmental contaminant," said Daniel Wartenberg of the Environmental and Occupational Health Sciences Institute in New Jersey. "You don't know the population risk, you don't know how long an exposure was . . . or maybe even what the [contaminant] was." The burden of investigating disease clusters falls on the state Bureau of Environmental Health Assessment, an office with an annual state budget of about $900,000 and about 50 full-time people. It must balance those requests with other important work on childhood lead poisoning and a slew of other programs. In recent years, the bureau's budget has been sliced by 40 percent.

In Belmont, Taylor's case is emblematic of both the hope and the frustrations of the hunt for disease clusters. Residents there have long complained of pollution from a metal-finishing plant called Purecoat North LLC. In recent years, the Environmental Protection Agency fined the business -- formerly known as Cambridge Plating Co. -- more than $130,000 for hazardous waste and federal air violations. A company official did not return a phone call from the Globe.

Residents around the plant say they have worried about their health for years, but it wasn't until recently that they started to believe all the reports of illnesses might be linked. "You have to wonder," said Audrey DiGiovanni, 39, who lives down the street from the plant and has three children. Her 74-year-old aunt, who lives on the same street, had breast cancer and her aunt's husband had prostate cancer. "We counted 13 people on this street alone who had cancer."

By painstakingly collecting anecdotes and mounting a case, Taylor and her neighbors were instrumental in getting state legislators to set aside $50,000 for an investigation by the bureau.

Now, bureau staff members are immersed in the first level of investigation: Trying to determine whether Taylor's neighborhood has actually sustained an unusually high level of disease.

It is not an easy task. Researchers know that at least one in three people will experience cancer in their lifetime, so having several people on a block with cancer may not be as unusual as most people believe. And cancer is far from a homogeneous disease, so two cases of cancer -- such as leukemia and brain cancer -- are likely to have different explanations.

Even if investigators do find a higher-than-expected rate of sickness, statistics suggest that what look like "clusters" often can be explained by simple chance. Even if cancer rates are normal, there are bound to be some neighborhood clumps of the disease -- just as a handful of confetti dropped randomly on the ground will fall more densely in some areas than others. As soon as an investigation is begun, researchers also look into what possible toxins might be contributing to the cancer cases -- if it turns out there is a cluster. To do this, they must look through old reports to document every type of chemical exposure that may have affected residents.

"It's hard to make those links, but we do sometimes succeed," said Suzanne Condon, state assistant commissioner of environmental heath.

As an example of success, she cited the "Civil Action" case, in which Condon's office found that leukemia rates in a neighborhood in Woburn were unusually high. Condon saidher office also developed an accurate model that predicted how contaminants flowed into wells that provided drinking water for pregnant mothers whose children became ill.

Since then, the office has investigated scores of potential clusters, and has published several reports that draw a link from disease to an environmental contaminant. But even those cases have been controversial: For instance, the state found elevated cases of leukemia near the Pilgrim Nuclear Power Plant in Plymouth in the 1980s, but a federal study looked at a broader geographical area and found no elevation.

Some critics of the disease-cluster idea say there have been no direct links shown at all -- even in the high-profile Woburn case.

According to Bill Thilly, a professor at the Massachusetts Institute of Technology, the Woburn cluster of leukemia could be the result of simple chance. He and a former student who has since received her doctorate conducted a study of mortality rates of 26 cancers in Massachusetts and several other states and found they all were in distributions expected by chance alone.

To explain his point, he uses a deck of cards: Chances of turning up two aces from the top of the deck "are pretty rare," he said, "but it happens. And it doesn't mean there is anything funny about the deck."

Epidemiologists dismiss this argument. They acknowledge that chance could explain away some clusters. But it doesn't prove that contaminants didn't cause the cluster, they say -- so public health departments still have a responsibility to investigate. They also complain that the state hasn't set aside enough money for investigations -- forcing residents like Taylor to take a political route to get studies funded, going to state legislators to get money set aside in the budget for health studies.

David Ozonoff, professor of environmental health at Boston University, said he believes there is a disincentive for local and state health departments to investigate clusters. "What if you say there is a cluster? . . . You are going to bring the mayor and selectmen on you; you are going to be labeled a cancer community."

For the residents at the center of such investigations, the chance of some kind of answer makes the fight worthwhile -- even if it means years of persistence and vast uncertainty.

Liz Lombard, who has scleroderma, helped spearhead a massive state study to find out why her South Boston neighborhood has four times the expected cases of the deadly autoimmune disease.

"My eyebrows are gone, my hands are like melted wax," Lombard said. "If we all just moved out of sight, it doesn't solve anything. We have to stay here and see it through so it doesn't happen to anyone else."

Beth Daley can be reached by e-mail at bdaley@globe.com.