Cancer Wars

An abnormal number of kids in Sierra Vista are getting leukemia. Why does the government insist that it's probably just a coincidence?

One Sunday night in October 2001, Dale and Kelley Durkit took their 2-year-old daughter Jessica to the emergency room at Sierra Vista Regional Center for festering spider bites on her foot.

The same brown recluse spider had apparently bitten her several times. They had just been at church, where someone had told Kelley about a good pediatrician. They'd been reluctant to have the bites treated by the pediatrician she had been seeing. He'd been dismissing their worries about the little girl's bruising and listlessness for months. Fortunately, the new pediatrician agreed to meet them at the medical center, and they took her in.

Sitting at the family's kitchen table two years later, Kelley winces as she recalls the last few hours before their lives changed. "The doctor looked at her and said, 'There's something very wrong with her. She needs to go to Tucson, to UMC (University Medical Center), tonight. Right now.'"

"I rode up with her, and Dale took Cody (Jessica's older brother) home and stayed with him," Kelley says. "At the hospital, they put her in this crib like a metal cage. I stood there rubbing her back until I couldn't stand any more and lay down on the cot." (Kelley was six months pregnant.)

"And then, it seemed like the next minute, it was morning, and there were three or four doctors in the room and one was saying, 'Your daughter has leukemia.'"

Kelley continues. "I said, 'You must be joking.' And he said, 'Believe me, I wouldn't joke about something like this.'

"I fell apart."

SIERRA VISTA HAS A CANCER cluster. Children are getting sick in the booming military town 70 miles southeast of Tucson, and nobody really knows why. Since 1997, 12 Sierra Vista children have been diagnosed with acute childhood leukemia, about 8 more than expected in a total population of about 40,000. One of those children is dead; the others are in treatment or in remission. The town's leukemia rate is three times what it should be.

Childhood leukemia is a bad, bad disease. It's an insidious cancer of the bone marrow and blood that's 99 percent fatal if left untreated. It prefers toddlers and preschoolers. The treatment, which saves seven or eight children out of 10, is a hellish, drawn-out course of chemotherapy that brings the young victims near death and keeps them there for months. Once the disease is in remission, patients are not declared clear of it until five years have passed.

Although leukemia has been studied intensively, cancer in small children--like most cancer--remains a medical mystery. But there's some hope that science will eventually solve the puzzle of leukemia. The fact that it tends to occur in clusters suggests that environmental factors increase the risk for the disease; the victims' youth means that their exposure histories are much less complicated than those of adults. Still, until someone figures out what's causing the outbreak in this high-desert town in the shadow of the Huachuca mountains, nothing can be done to reduce its children's risk.

Since the cluster was confirmed by the Arizona Department of Health Services and reported to the national Centers for Disease Control in Atlanta in 2001, government agencies have repeatedly declined to investigate, and have often seemed reluctant to discuss the cluster--or even to acknowledge its existence.

This stance is by no means peculiar to Arizona. Government science's approach to cancer clusters is essentially defensive. And its methods, which have not changed substantially in 30 years, seem to have been designed expressly to make outbreaks vanish into clouds of case exclusions and probability calculations.

While the state health department and the CDC have resisted calls to investigate, two researchers from the University of Arizona--Mark Witten, a toxicologist and professor of pediatrics, and Paul Sheppard, a professor in the UA's distinguished tree-ring lab--have been doing it on their own, in their own way. For more than two years, they've followed clues that appear to link the cluster to those in three other towns: All four towns have a lot of leukemia and an unusual and rising abundance of tungsten, a heavy metal whose biological effects are all but unknown. While trying to find out whether this association of tungsten and cancer is significant, Sheppard and Witten have gathered data on weekends and during vacations, logging thousands of miles between Tucson and the sites they're investigating. Witten says he's spent at least $10,000 on travel and analysis of air, soil, water and tree-ring samples. That was money he was paid for working on an aerosol measles-vaccine project sponsored by UNICEF.

The shoestring phase of their project is now over. Sheppard's presentation about it to the American Geophysical Society in December was well-received and featured in Science Update, a publication of the respected science journal Nature. In December, they received substantial grants from the Gerber Foundation and the Cancer Research and Prevention Foundation to continue and expand their research.

For its part, the ADHS has declined to do the biological testing in Sierra Vista that Witten has asked for, testing that's necessary to correlate toxin loads in residents' bodies to toxins in the environment. According to Michael Murphy, spokesman for the state health department, the state is "more than happy to support research that is scientifically valid," but has not funded nor assisted the UA researchers. He refuses to explain further.

Much of the Sierra Vista business community prefers that the cluster receive as little mention as possible, and is wary of the UA scientists. According to a report in the Sierra Vista Herald, members of a county-wide business group, Team Cochise, concluded a meeting last August by agreeing that "Witten is using anything he can to obtain a large federal grant."

People who make their livings in Sierra Vista have every reason to be tense about bad news. The timing of the cluster could not be worse. The town is facing the possibility of losing its economic engine, Fort Huachuca, in the next round of Pentagon base closings. Everyone's worried, and nobody wants leukemia to be part of the city's conversation with the Army. Many townspeople would prefer that the leukemia had never been mentioned, much less publicly discussed by someone as voluble and PR-savvy as Witten.

Then there's the passionate minority that's frustrated with the government because of what it sees as a languid response to a pressing problem. They like it that Witten gives presentations about what he's doing, and they like it that he's trying to find answers.

This polarization is typical of what goes on around clusters. The politics of cancer are fierce and surprisingly subjective--every fact and assumption, even the very idea of cancer clusters, is in question. This is nothing like Hollywood science. In this world, scientists disagree about the facts, the methods, the conclusions and about what the public should be allowed to know.

Last week, the CDC announced that it will visit Sierra Vista before the end of the month. At the same time, in a letter responding to Rep. Jim Kolbe's second formal request for help, CDC Director Julie Gerberding made it clear that any action the federal agency takes will be the ADHS' call. A document from the National Center for Environmental Health says: "The number of children diagnosed with leukemia, although a tragedy for each family, is still fairly small and may not in fact constitute an elevated rate or true 'cluster.'"

Welcome to the cancer wars.

Childhood leukemia crept up on Sierra Vista. According the Arizona cancer registry, which gets its figures from the department of hematology/oncology at the University Medical Center, not one child in the area was diagnosed with childhood leukemia in 1995 or 1996 (see chronology). The years 1997, 1998 and 1999 saw one new case per year. From a public health point of view, this was nothing to be alarmed about: Three cancers in five years is exactly the expected incidence of the disease in a population of roughly 10,000 children like Sierra Vista's.

In 2000, two more children turned up sick, bringing the total to five--four, according to the state health department's cancer registry--which pushed the incidence rate above average. (The uncounted case was Annastacia Warneke, 6, whose family had moved away two weeks before she was diagnosed. CDC reporting rules meant that hers was not counted as a Sierra Vista case. Her father, Mark Warneke, knew of the other cases and protested, triggering one of the cluster's first controversies.) The disease rate was still not high enough, though, to declare a cluster.

In 2001, three more cases were diagnosed: One was a 9-month-old baby who died. At this point, the local paper, the Herald, began reporting each new case, much to the dismay of readers who were concerned that the stories were bad for the town. (The Herald is owned by Wick Communications, the same company that owns the Tucson Weekly.)

In June of that year, a Tucson woman whose son was getting chemotherapy at UMC told Witten that a number of the kids being treated there were from Sierra Vista. Witten checked it out. He says that he called Dr. Tim Flood, director of the state cancer registry, to tell him that there was a leukemia cluster in Sierra Vista. In September, Flood confirmed that the town had a "borderline cancer cluster," and that one more case would push it over the edge. Channel 13 ran a story on it some weeks later.

Dale Durkit saw the segment on the TV in his daughter's room at UMC. "I went out and called Channel 13 from the phone in the hall and told them that I had their seventh case, and they could come see her, if they wanted," says Durkit.

The station sent a crew over to the hospital and filmed a segment on Jessica Durkit and her family.

"You wouldn't believe how fast the administrators had me in their office," says Kelley. "They wanted to know how the TV people got into the room. I told them that my husband invited them. And then it was, 'You don't want those people bothering you, now do you?' I said they didn't bother me at all."

One more child was diagnosed in 2002. The family had moved away before the leukemia was identified, so the state cancer registry did not include it in its tally. The exclusion made activists like the Durkits even more convinced that the state was trying to downplay the outbreak.

In 2003, the cluster became undeniable when three more children were diagnosed, two of them within a month. Suddenly, leukemia news was everywhere. In June, Dawn Harvey, a 23-year-old who was too old to be counted, died. (The cut-off age for registered cases of childhood leukemia is 14. The explanation given for this is that after 14, kids may be engaging in sex and taking drugs, which makes the data unusable.)

In August, there was a fund-raiser for 13-year-old Susan Taylor of Hereford. Chemotherapy had damaged her heart valves so badly that she needed open-heart surgery to repair them.

THE QUESTION OF WHETHER THE state health department would investigate first arose in 2001, when the cluster became official. The answer--"no"--took a while to emerge.

In September 2001, Flood said that he was waiting for more information before taking action. In September 2002, following the announcement of a new case, Kolbe made his first formal request to the CDC for a full investigation. The CDC declined, saying that the state was capable of doing any study that was needed. Flood said that the cluster was still "marginally statistically significant," and that his office was waiting for the results of the CDC's ongoing investigation of a similar leukemia cluster in Nevada.

Fallon, Nev., is a small military town 60 miles outside Reno with two claims to fame: It's home to the famous "Top Gun" training base for Navy pilots, and it's home to the worst per-capita leukemia cluster ever recorded in the United States--16 cases in seven years in a total population of just 8,000. (Three months ago. another case was diagnosed just outside the Churchill County line. The county is the cluster area as defined by the CDC, so the new case has not been added to the Fallon count.) In addition to its simultaneous outbreak of leukemia, the town resembles Sierra Vista in several suggestive ways: It has a military airfield, a high-desert climate and a close proximity to old tungsten mines. And, like in Sierra Vista, the government doesn't know what to do about the leukemia cluster.

In early statements about Sierra Vista, Flood emphasized the consistent failure of CDC-style epidemiological studies to establish links between contaminants and cancer. One reason he and other agency scientists cite for not tackling cluster investigations is simple discouragement: The 108 full-on cluster investigations conducted by the CDC have not uncovered a single link between cancer and environmental factors. There are two possible explanations for this remarkable run of negative outcomes: There are no links between the environment and cancer, or government-sanctioned investigation methods don't work.

One of the last and most frustrating--some would say infamous--of these studies took place in Maryvale, an industrial neighborhood in West Phoenix that had seen 49 leukemia cases in a 21-year period beginning in 1965. (For much of the duration of the cluster, there was no treatment for leukemia. Virtually every child diagnosed died.) The investigation was conducted by the ADHS under the oversight of the CDC and began in 1987--after local media started covering the cluster and criticizing the health department's failure to act. The media also discovered that in 1982, the agency had told the parochial school principal who first noticed that children were dying not to talk about it.

The Maryvale study, carried out with the oversight of the CDC, cost millions of dollars and went on for more than 10 years. It finally found no link between environmental factors and the leukemia in the area. This wholly negative result was surprising to many observers: Even before the cluster was identified, the city of Phoenix had begun closing the neighborhood's trichloroethylene-tainted wells, and in 1987, the whole area was designated a Superfund cleanup site. According to its critics, the ADHS study did not look seriously at the water supply, but busied itself with collecting data on a thousand variables--including diet and TV ownership. When the results were released, the remaining neighbors understandably saw this as one arm of the government acknowledging a health threat while another, charged with ensuring public health, saw none. Most of the lawsuits brought by survivors were unsuccessful, and few of the companies that had dumped chemicals were even fined. Some cynics believed that the study had been drawn out and out with the express intention of deflecting litigation against the city and the polluting industries.

BACK TO SIERRA VISTA. IN December 2002, the CDC's Gerberding turned down Kolbe's request for a federal study, explaining that "the Arizona Department of Health Services is equipped to handle the investigation." To date, the state has done no investigating beyond a review of the records of local water companies and industries.

"There are limits in what we can do," explains State Epidemiologist Dr. Bob England. "We've taken the appropriate steps to identify environmental problems that we know are problems. We checked for known carcinogens in the water and air and didn't find any. I know people in Sierra Vista are frustrated, but no one could be more frustrated than I am by this situation."

In February 2003, the CDC's expert report on the Fallon cluster was released. The CDC had gone into Fallon under intense political pressure: The Reno Gazette-Journal began aggressively covering the cluster in 2000, and in 2001, Sens. Harry Reid, D-Nev. and Hillary Clinton, D-N.Y., held a Senate committee hearing about it.

Once again, the CDC failed to find any statistical correlation between leukemia cases and the environment. What it had found were high rates of arsenic and tungsten in people's bodies. Long-term arsenic exposure has been linked to many cancers and other serious diseases, but is not known to cause leukemia. The final report advised residents to drink bottled water to reduce their arsenic exposure.

Tungsten levels were even higher. Tungsten is a heavy metal with a high melting point that's used to make light-bulb filaments, jet engines, tools and nosecones for bombs and rockets. Its ores have been mined near both Fallon--where there used to be a tungsten processing plant as well--and in the Huachuca mountains outside Sierra Vista. Very little is known about its biological effect, and no safe level of human exposure has been determined.) Urine tests showed that Fallon leukemia patients had average levels of tungsten 18 times the U.S. norm, and twice as high as Fallon adults. But after putting the data through a logistic regression analysis designed to find patterns that are not due to chance, the expert panel found no correlation between the town's leukemia excess and tungsten, or any other environmental factor.

Based on the CDC's raw data, Witten and Sheppard intensified their tree-ring testing around Fallon and found that environmental levels of tungsten were high--and more significantly, increasing. Tree cores taken around Sierra Vista showed the same pattern, as did those from two other towns with clusters: Elk Grove, Calif., a suburb of Sacramento, and Hoisington, Kan. No one knows why there's more tungsten now than there used to be in any of these places. Witten and Sheppard suspect that drought could have something to do with what's happening in Fallon and Sierra Vista--the towns may be dustier than they were, meaning that people would be breathing in more of whatever's in the soil.

"But that's just a guess, one of the possibilities," says Witten. "What we know for sure is that trees have been taking up more tungsten in the last five years than they did 20 years ago. And there's been more leukemia in last five years. What are the odds of that happening by chance in four different places? It seems worth looking into."

In spite of its conclusion that there was no demonstrable connection between tungsten and cancer in Fallon, the CDC panel called for immediate study of the metal's toxicity, plus a comparison of Fallon to other Nevada towns with high tungsten levels.

So the Agency for Toxic Substances and Disease Registry, an agency of the U.S. Department of Health and Human Services, went out and gathered data on three other Nevada towns with old tungsten mines. It released its findings in July of last year. In two of the towns the agency studied, researchers found elevated tungsten levels in people's urine, but nothing like the Fallon levels. And no leukemia cluster. In the third town--tiny Yerington, Nev. --the researchers found very high levels of tungsten in the water and in residents' bodies, although not as much as in Fallon--and, once again, no leukemia cluster. The CDC's conclusion from these findings was that tungsten "is not a concern for cancer."

Witten contends that this conclusion is unwarranted.

"There is no way, based on those numbers, that you can rule out tungsten," he maintains. "Eight times the normal level of tungsten, which is what they found in the Yerington kids, is not the same as 18 times, which is what the Fallon kids have.

"Look, one of the main things toxicologists look at is dose response level," he says. "A substance can be harmful at one dose, but poisonous at higher one." This is why the package tells you to take two Tylenol at a time, not 20.

"Another factor is whether the body can get rid of a particular toxin. We know that other heavy metals like lead and mercury are cumulative poisons," Witten says. "You also need to remember that kids' metabolisms run three times faster than ours. Their bodies process toxins differently.

"And when they talk about so many times the standard level of tungsten? That standard is just an average, not a desirable or safe level. We have no idea how much tungsten exposure is safe."

Witten and Sheppard continue to focus their investigation on tungsten, although they haven't ruled out jet fuel--on which Witten is an expert--or anything else. (They've found JP-8 fuel in Fallon trees, but the CDC did not do the standard test for exposure to JP-8 in Fallon, so they cannot compare environmental levels to biological ones.)

"I suspect that tungsten is actually a co-factor: Exposure to it is just one contributor to the development of leukemia in these kids. But I have a strong hunch that it's significant," Witten says. "For one thing, it's the only exposure besides arsenic that pops out of the Fallon data. Why not try to figure out if it plays a role? If it turns out that it does, we may be able to tell people how to reduce their kids' risk. If it's in the soil, say, then people would know to keep their kids in on windy days and vacuum a lot.

"But even if tungsten turns out to be a dead end," he says, "we will have eliminated it as a cause, and that's one more piece of information about the disease. We've got children dying. Why not pursue the clues we've got?"

The response of the state to the Fallon report was quite different.

"We were really hoping that something solid would come out of that study," says England. "If the CDC had found that tungsten was a possible cause, then we might have had something to look at."

Citing the government's dismal track record on clusters, England says that he is reluctant to start an epidemiological study that he is certain will lead to nothing but frightened people and depressed property values.

But when asked whether he thinks that the CDC's methods could identify a causal factor even if there is one, he pauses. "I don't know."

THERE SEEM TO BE TWO entirely different kinds of science operating here. The CDC, which sets the rules for how such studies are conducted by the states, deals with evidence in a highly abstract and bounded way. From the outside, it looks as if government scientists don't really want to find environmental causes. According to England, that's because they know they can't find them. He feels that the state agency's most important role at this point is to educate people about leukemia clusters and to work with the CDC on getting Sierra Vista into a larger, more promising study.

"So far, we have done a bad job of communicating with people in Sierra Vista," England admits. "We have really screwed up there, and I'm looking forward to changing that."

Witten and Sheppard's approach--and that of their new colleagues, Deborah Frisch, a retired program director for the National Science Foundation and an expert in risk analysis, and Eric Betterton of the UA's atmospheric sciences department--is based on the assumption that there is something to be found, and that they may be able to find it. Witten first went into the investigation looking for jet fuel, but partially changed tack when he saw the CDC's numbers on tungsten in Fallon. To his critics, this is a sign that he is rooting around looking for anything and altering his hypothesis as he goes along.

He agrees that that's how he and Sheppard started out. "That's where science starts. How are you going to find anything if you don't go in with an open mind?"

He and Sheppard are pursuing clues suggested by preliminary data and analyses, and developing new techniques as they go. Sheppard is pioneering a new use for tree-ring analysis--using dendrochemistry to establish histories of environmental change related to changes in health. Witten, using tools not available even five years ago, has learned that tungsten increases expression of a gene associated both with leukemia and with rhabdomyosarcoma, an extremely rare muscle cancer that has also been turning up in all four towns they're looking at.

He's also found that exposure to tungsten encourages cultured cancer cells to grow faster, and one experiment showed that while lab animals excrete most of the tungsten they're fed, what they retain is stored in their bone marrow--where leukemia originates. These are only hints--scientists have to replicate findings before they're useful--but they suggest that he and his colleagues may be on the right track.

"As a scientist, I find it very hard to understand the agencies' lack of curiosity," says Sheppard. "I understand on one level, of course, that there are lots of reasons for them to want to stay out of it, but as a scientist, I can't fully grasp it. Of course we're driven by the thought of children dying--we have kids ourselves. But on top of that, this is very exciting research."

Adds Witten, "I don't get it, either. I ask the health department and the CDC to go in and get the kids in Sierra Vista to pee in a cup so we can see what toxins they're carrying around in their bodies, and they won't. Why? Because there's no evidence of a known environmental cause in the environment, according to their standards. Well, that's kind of circular, isn't it?"

The researchers will need more grant money to do blood and urine draws in Sierra Vista. The grants they have now are paying for them to expand their environmental testing and to do mouse-model studies on tungsten exposure in mice.

Frisch, who has spent time in the federal science establishment, elaborates: "It's clear that the case study approach doesn't work. There's an opportunity now to look at multiple locations and multiple types of cancer. You'd think they'd jump at the chance to conduct a real scientific study.

"The Nevada tungsten study is junk science," she says. "It only proves that tungsten alone doesn't cause leukemia. It doesn't rule out tungsten as a contributing factor. It doesn't look like the government is interested in finding an answer."

THERE'S NO DOUBT THAT government agencies do all sorts of things to avoid looking for or finding environmental causes of cancer. They throw out cases, limit samples, accept industry testimony without question, use fantastically complex statistical methods and write conclusions that don't always match their data. A close look at the Fallon investigation shows that every assumption, every decision, leans toward: "There's nothing here. Don't worry; be happy."

A final report on Fallon by the Agency for Toxic Substances and Disease Registry last July, for example, concludes that "there are no past, current or future public health hazards from exposures to Naval Air Station Fallon substances in the environment" (emphasis added). It's difficult for the layman to understand how anyone, under any circumstances, can predict the future.

Several factors seem to contribute to government scientists' drive to maintain the status quo. One is money. Epidemiological investigations are expensive. Another is politics. These are people who, after all, work for the government. The wealthiest and most influential people in any given area are utility owners, industrialists, landowners and business people. They don't want trouble and they have clout. For example, Sierra Vista is a fast-growing, ambitious town where many prominent citizens have bet on continued growth, and don't want to hear about children with cancer.

"Of course we see those headlines and cringe," says Judy Gignac, vice president of a local water company and general manager of a land-development concern. "But I'll tell you, nobody would like this resolved more than me. Yes, because of business, but also because of those kids."

There's also the messiness and work and expense of cluster investigations, and the public anger and criticism that comes at the end when they're unsuccessful. (Which is almost always. One of the very rare examples of a successful cluster investigation by the government was the identification of HIV, the virus that can lead to AIDS, as the ultimate cause of an outbreak of Karposi's sarcoma in the '80s.)

"We go in and get people's hopes of an answer up, and then let them down when we can't find it," says Dr. England. "I hate doing that to people."

Another factor explaining the government scientists' drive to maintain the status quo is belief. Many epidemiologists and physicians--and for obvious reasons, people involved in polluting industries--do not think that there are such things as "real" clusters at all, except for certain occupational instances. They believe that all residential clusters are statistical blips in a random distribution.

Random cancer clumps are everywhere. As England explains it, roughly one in every 20 randomly chosen areas will have an excess of cancer that's entirely due to chance.

Still, one feature of the Fallon and Sierra Vista clusters makes them look frighteningly real: In both places, children have continued to get sick after the cluster was declared. In Sierra Vista, the cluster has actually accelerated since it was confirmed in the fall of 2001.

To see why this is significant, you have to understand the skeptics' central argument. The problem they most often point to is that the boundaries of clusters are defined by the cases. You notice five cancers on your block, decide that the block has a cancer cluster, and sure enough, you've got a cluster inside the boundary you've drawn. But what you don't notice is that there are some blocks around yours that have no cancer, and others that have a normal number, so the neighborhood rate is somewhere near average. Your cluster is not "real"--it's just an unfortunate illustration of the fact that randomness is lumpy. Cluster skeptics believe that all clusters are like this. There's a reason why "cluster" is in quotes in latest document from the CDC, cited previously.

They illustrate the argument with the Texas Sharpshooter Fallacy, which goes like this: A stranger claiming to be a dead-on shot, able to hit any target, "proves" his skill by taking a random shot at the side of a barn. Then he goes and draws a target around the bullet hole, points out that his bullet is right in the middle of the target, and declares that this proves he's a sharpshooter. In this analogy, the bullet corresponds to a group of cancer cases; the circle around the hole is the area and period of time in which the cases appeared. The cluster is bogus.

But when cases keep appearing after the boundary is set, as they have in both Fallon and Sierra Vista, the story gets another act. The gunman fires his first bullet (the first seven Sierra Vista cases) and draws his circle around it (the cluster is confirmed). But then he goes back out to where he was, fires and hits the target he's drawn (another five cases appear). Because the odds of hitting the same place again by sheer chance are astronomical, it's clear that the man is a sharpshooter after all. The cluster is real.

England concedes this insofar as to say, "The fact that cases continue to appear does make the cluster more credible."

TO THE PARENTS OF LEUKEMIA victims, statistical arguments about whether the cluster is real or not are at best irrelevant, and at worst, infuriating. They've watched their children lose their hair, vomit constantly, submit to innumerable blood tests and marrow biopsies. They hear about other sick kids. They ask about causes and get explanations of random distribution. When they continue to wonder why their child got sick, they're told to put the whole thing behind them.

Dale Durkit believes that the cluster will continue, and worries about the kids of Sierra Vista. (Today, as you saw on the cover, Jessica is a lively, charming 5-year-old two years into remission.) He and Kelley will tell you that they made it through Jessica's harrowing eight-month treatment only with the help of their strong evangelical faith and the support of family, their church and Dale's co-workers at Fort Huachuca, where he is a medium-equipment operator.

Kelley gave birth halfway through Jessica's first course of treatment at UMC in a room six floors above her daughter's. The family would have lost its modest home had Dale's parents not intervened. Dale did the 190-mile round trip to UMC every weekday for months so he could be with his family and still keep his job. They're terrified every time Jessica runs a slight fever, and they still have bills in collection.

"They tell you your child has leukemia and your life is chaos, just like that." Kelley says.

Dale obsesses over the cause of the outbreak--he believes it's linked to jet-fuel exhaust. (Planes from the fort turn around directly over their mobile home, and the machines he operates at work run on JP-8.) Both of them wonder why people in the town are so quiet.

"Children here have been misdiagnosed over and over," Kelley says. "When we'd taken Jessica to the doctor about her bruises and tiredness and not eating, he kept saying, 'Oh, she's just 2.' Don't worry.'" One nurse who saw the child's ugly bruises lectured Kelley about child abuse.

"But I've heard," Kelley says, "that our old pediatrician is doing things very differently now."

Dale is still anxious: "I read the other day that there's an epidemic of child abuse in Cochise County," he says. "That really scared me. I'm afraid that some of those kids who are bruised up have leukemia, and no one realizes they're sick.

"People here need to know about this, if only so they can watch their kids," he says.