I knew that the individual papers must tell a mixed story. The purpose of the meta-analysis was to combine the results of different studies to make sense of that confusion. I spent the next week immersed in books and papers on the methods of meta-analysis, picking the brains of Chalmers and Mosteller and planning how to approach the challenge of combining the results of the studies. The Boston sky was often dark by the time I emerged from the windowless confines of our basement warren at the end of each workday. On those nights when the cafeteria was still open, I would sit and eat while reading a few more papers before climbing up to my small room on the third floor.
With nothing more than a bed, a chair, a desk, a lamp, and an alarm clock, the room suited the monklike existence I would lead during the two months I had to complete the project. Each night I plowed through a few more papers before collapsing into bed. Each morning I woke early for a run along the Fenway and a quick breakfast before crossing Longfellow Avenue, a narrow canyon between the ever-growing cliffs of Harvard’s medical empire. The early morning was my only chance to steal a quiet moment for work, so I made a point of arriving at the office before the crowd came to bemoan the latest pummeling of the Red Sox in the American League play-offs.
As the play-offs wore on, I worked with Bruce to make sure we had not missed any papers. Bruce was among the most literate people I had ever met. He once told me that he read some sixty books each month and was rarely without a book in his hand. Bruce was happiest in the musty stacks of the world’s great libraries and archives in search of hidden treasure. Together we exhausted Harvard’s Countway Library, tracking down every possible article.
The studies all asked the same question, but took widely different approaches to answering it. The challenge of meta-analysis is finding the common threads in the disparate analyses and using those threads to stitch them together in a meaningful way. Tom’s office was a beehive of intellectual activity, but not always a great place to solve a puzzle. At one point, as I struggled to make the study work, I went to Fred for help. I told him I needed a place to think.
Fred’s office was an orderly refuge amid the creative chaos of the Technology Assessment Group. He and Tom were the intellectual odd couple, two men with opposite personalities who had found common ground in the late autumn of remarkable careers. A man of few words, Fred hesitated a moment before opening his desk and handing me a key. “Why don’t you use the hideaway,” he said. This was a key to the world of the chaired professor, a world where one finds space in a place where there is seemingly none to be had.
I followed Fred’s directions through the School of Public Health to a nondescript door that I had passed a dozen times without noticing it was there. Behind it was a large room, twice the size of most offices in the school. The walls were lined with bookshelves filled with carefully organized bound copies of the world’s major statistical and biostatical journals going back several decades. In its center were the hideaway’s only pieces of furniture: a small table and a wooden chair. On the table sat a neat stack of yellow legal pads and a box of pencils. Nothing else. I felt as if I had entered some sacred temple of the intellect.
There, in Fred’s hideaway, I put together the pieces of the puzzle. As I assembled the data that Tom and one of his graduate students had pulled from the papers, a consistent picture began to emerge. A lifetime of drinking chlorinated water increased the risk of bladder cancer. It appeared that it might increase the risk of other types of cancer as well, especially colorectal cancer.
I got a hint as to the controversy I was about to stir up when I began to send the paper out for publication. I routinely received two responses: one would praise the paper and herald its scientific import and the other would attack it as an inappropriate use of meta-analysis. The editors, looking for two strong reviews, would reject it. In the summer of 1991, I found a way to get it published with the help of a hurricane named Bob and a remarkable woman named Devra Davis.
That spring I had been invited to serve as an adviser to a National Academy of Sciences Committee on Environmental Epidemiology. Devra Davis, a longtime advocate on issues related to environmental health, was then working for the National Academy and was in charge of shepherding the committee through the process of producing a report. The committee had already met several times in Washington when Devra scheduled a meeting at one of their other conference facilities—an old Cape Cod estate near Woods Hole. As I arrived at Logan airport, I noticed the headlines in the Boston Herald. “Bracing for Bob,” it read in the paper’s usual hyperbolic type.
Twenty-four hours later I huddled in our hotel with several of the committee members. The air cracked with the sound of huge old trees splitting into pieces. Across the street the storm surge rolled in, ripping boats from their moorings. The devastation went on for hours, interrupted only by a few ethereal minutes of calm as the hurricane’s eye passed over our heads. With power lines down all over the Cape, the arrival of night plunged us into complete darkness. We awoke the next morning to find the shoreline littered with boats and the roads clogged with tree limbs. The power would not return for three days. Hurricane Bob was one of the worst storms ever to hit the Northeast.
The committee had no choice but to make do. We met in unlit rooms, ate food warmed over Sterno, and went to bed by candlelight. One day I presented my study to the committee and explained my difficulty in getting it published. Devra suggested I send it back to the American Journal of Public Health. The journal, she said, had a new editor who might reconsider a decision by his predecessor to reject it. I took her advice and the following summer it published my paper.
The reaction shocked me. Within twenty-four hours the National Cancer Institute (NCI) and the Environmental Protection Agency issued a press release denouncing the study. The NCI response surprised me since their own scientists had studied the risks of chlorination by-products with results similar to mine. I suspected that I had treaded on the turf of the EPA, but again I could not understand the vehemence of their response. It would take years before I could understand the strange world of science, politics, and money that I had just entered.
In August 1992 I received a call from a representative of the International Life Sciences Institute (ILSI). They were hosting a meeting on the safety of drinking water disinfection in Washington, D.C., and hoped that I would attend. I had been aware of the meeting, but had only recently joined the faculty at the Medical College of Wisconsin and could not afford the trip. When I explained, they offered to pay my way. I agreed to come assuming that they wanted me to participate in a discussion of my paper. I could not have been more wrong.
The food and beverage industry, I would later learn, had established ILSI to study nutrition and food safety. The woman who had called to invite me to the meeting was an employee of Coca-Cola, a major supporter of ILSI. When I arrived at the meeting, I noticed immediately that I did not have a place on the panel of speakers dealing with chlorination by-products. I concluded they must have flown me to Washington to comment from the audience on issues raised by the speakers.
On the second day of the meeting, I found a seat in the vast, generic meeting hall, which was filled with hundreds of engineers, microbiologists, and toxicologists. I listened to one speaker after another, waiting for my chance to talk. Several speakers commented on my study, but none of them offered any specific critique. The first speaker to directly address the epidemiology of disinfection by-products was Gunther Craun, a former EPA employee who had gone on to become a successful consultant to the drinking water industry. He offered his own review of the epidemiological studies on chlorination by-products, concluding that they were inconclusive with respect to the risk of cancer. In his closing comments, he dismissed my study as an inappropriate use of meta-analysis. When Craun finished speaking, I stepped to the microphone reserved for questions.
“You have criticized the use of a meta-analysis in this context. A meta-analysis combines the results of studies using a strict, objective, quanti
tative scheme to draw conclusions from the literature. A review of the literature, the only alternative, relies on unspecified, subjective, qualitative scheme to draw conclusions. Is a qualitative, subjective method really superior to a quantitative, objective method?”
Craun simple restated his belief that the meta-analysis should not have been attempted. I started to ask if he considered it appropriate to reject meta-analysis in favor of a literature review, which offers no protection against the biases of the reviewer. As I did the panel moderator cut me off. There was, he said, no time for more questions.
If I had any doubt about the agenda of the EPA, it disappeared as I listened to a speaker on the next panel, Pat Murphy, an epidemiologist for the drinking water office of the EPA. Murphy spent almost her entire time at the podium on an extended critique of the meta-analysis. Murphy, it turned out, had played a key role in an internal EPA review that had evaluated the potential for some sort of meta-analysis of the studies on chlorination by-products. She had concluded that a meta-analysis was not feasible. “The Morris meta-analysis,” she explained, “combines apples and oranges and comes up with a fruit salad.” She smiled at her own joke.
Had Murphy been speaking to an audience of epidemiologists, the flaws in her critique would have been obvious. However, I believe I was the only epidemiologist in the audience and I know I was the only physician. My experience with Gunther Craun had made it clear that I would have no opportunity to respond in any meaningful way to her speech. It seemed that I had been invited to be seen, not heard.
I was getting my first glimpse of the strange interface between the EPA and the drinking water industry, and I was beginning to realize just how many toes I had stepped on. Without intending to I had indicted the EPA drinking water research laboratory, which had signed off on the safety of our water supply under the existing standards. With a meta-analysis that had taken six months to prepare and write into a manuscript, I had also stolen the thunder from epidemiologists, including those at the NCI, who had spent years on the painstaking process of collecting the data from people all around the country to look for the connection between drinking water and cancer. I also learned that I had contradicted the conclusions of the world’s experts on chlorination by-products and cancer who had recently met at the International Agency for Research on Cancer (IARC) and concluded that there was not enough data to draw any conclusions about this relationship. They refused to even allow the possibility that the by-products were carcinogens based on the available data.
Above all, limitations on the use of chlorine could create a massive problem for the drinking water industry. Drinking water treatment is unlike other regulated industries such as the automotive industry, the power industry, or the chemical industry. Those industries are paid to produce products that are not directly related to the quality of the air or water. Pollution occurs as an unintended consequence of the manufacture of cars, power, and petrochemicals. Protecting the environment will not in general improve the quality of their product. Hence those industries have a relationship with regulators that is inherently adversarial. The drinking water industry, on the other hand, is in the business of cleaning water and selling it. Requiring them to produce cleaner water seems consistent with their mission. On the surface it appears that the EPA and the drinking water industry have the same agenda. But the story is more complex.
Drinking water treatment plant managers have two major goals. First and foremost they must prevent massive outbreaks of waterborne disease. This means disinfecting the water and chlorine is the standard method for accomplishing this. The second requirement is to keep their bosses happy. In most cases that boss is an elected official who is never happy with the thought of asking for more money from his constituents. Using less chlorine, adding systems to remove its by-products, or introducing new treatment technologies each threaten one of these core missions. Furthermore, removing by-products would yield improvements in water quality that their customers, the people paying for the change, might not even be able to detect. In other words the drinking water industry had a tremendous vested interest in my being wrong.
But I soon learned that the industry had other, more urgent problems. As I sat at the ILSI meeting waiting to hear from the epidemiologists, I listened to engineers and microbiologists describing the world of water. What I learned suggested that water suppliers faced a problem that could prove far more destructive than any possible cancer risk from chlorination by-products.
The threat took shape through the words of Joan Rose, a tall, blond microbiologist from the University of South Florida. Dr. Rose described the watery world of protozoa in stark terms. These microscopic creatures lie somewhere between bacteria and true animals. They cause some of the most deadly human diseases, including malaria and dysentery. Rose showed data demonstrating that protozoa routinely contaminated rivers and streams throughout the United States. She also showed that specific protozoa often made their way through water treatment plants and into America’s drinking water.
As she spoke I thought back to my first encounter with a protozoal infection. As a medical student, I had helped care for a Milwaukee policeman. He lay in a hospital bed as the parasite slowly consumed him. I might have expected him to recover from the infection, but for a blood transfusion he had received two years earlier. The blood contained HIV (human immunodeficiency virus). As a result he lacked an effective immune system and we had no effective antibiotic for treating his infection. So we could only offer him temporary support in a losing battle for his life.
The title of the ILSI meeting, “Safety of Water Disinfection: Balancing Chemical and Microbial Risk,” told the story. After decades of believing that water treatment had made major outbreaks of waterborne disease a thing of the past, the industry faced a growing crisis. On one hand mounting evidence suggested that chlorine, an essential ingredient in that success, posed serious, unforeseen risks. On the other work by Rose and other scientists raised the possibility that the corpse of waterborne disease was not as dead as they had hoped.
With images of that emaciated policeman in my head, I listened in stunned disbelief. If what Rose and others were saying was correct, the American water supply was a disaster waiting to happen. Disaster did not have long to wait.
10
SPRING IN MILWAUKEE
The winds of chance blow through the careers of every scientist, but few feel the effects of those winds more profoundly than epidemiologists. Proximity to disaster can define a reputation. Late in the spring of 1993, as my plane rose up over the blue collar landscape of Milwaukee’s south side, chance was on the move. I had been on the faculty of the Medical College of Wisconsin for two years. My father was dying of prostate cancer and I was flying east to be with him and my family. I stared down as the plane veered out over Lake Michigan. The water, normally a cold, dark blue was pale and brown. My thoughts were ahead, in Connecticut with my father. But my future was below me in those turbid waters.
March, as always, was a messy time in Milwaukee. The Bucks were limping toward the end of a dismal season, hopes were riding high on the Brewers with the approach of opening day, and the weather was toying with the city’s emotions, swinging wildly between spring and miserable. First came the snow. The second week of the month had dumped six inches on the city before clamping it in a vise of dry arctic air. Nighttime temperatures dipped to three degrees. For almost a week, the snow gleamed under an icy sun. Then warmer air arrived laden with moisture. Alternating waves of rain and snow left the city awash in slush. Snowmelt filled the storm sewers. The rivers grew fat and angry. The Menomonee and the Kinnickinnic roared into the Milwaukee, which sent a plume of brown water surging into Lake Michigan. The rains and runoff of March carried more than mud to Milwaukee. They brought the seeds of catastrophe.
Tuesday, March 30, was a brilliant, sunny day with a gentle breeze. It seemed, for a moment, that winter had left the house, but the next day, it stormed back in like an angry drunk. The wind spun to the wes
t and cold air roared off the plains. The skies opened and temperatures plummeted. What began as heavy rain turned to sleet. As April began, ice glazed the city and a bitter wind blew snow into every corner.
For Mark Rahn, a forty-four-year-old car salesman from Chicago, the April Fool’s Day snowfall was little more than scenery. As he walked slowly out of the Milwaukee County Medical Complex, he had other things on his mind. Mark Rahn was trying not to die.
Locked in a pitched battle with leukemia, he had come to Milwaukee to undergo an extreme form of treatment known as a bone marrow transplant. Drawn by the reputation of physicians at the Medical College of Wisconsin, he had subjected himself to radiation and chemotherapy intended to destroy his bone marrow. Then, with his own ability to generate blood cells gone, he received a small amount of healthy bone marrow cells from a donor. He would then need to wait for those cells to find their way to the hollows left by the destruction of his own marrow, set up house, and create new blood cells for him.
Half the battle in surviving leukemia is surviving the treatment, particularly when the treatment is a bone marrow transplant. The drugs and radiation had ravaged the healthy cells in his body. They had eroded the lining of his gut, leaving him vomiting and often unable to eat. As his red blood cells had died off, he had become anemic and needed transfusions to survive. Worst of all the elimination of his white blood cells had stripped him of his ability to fight off infections. A few fungal spores drifting on an air current, an invisible clump of viral particles carried in on a visitor’s hand, or even the bacteria from his own gut could kill him.
After weeks in an isolation unit breathing microfiltered air, things were looking up. The chemotherapy had stolen his thick head of hair, the steroids had bloated his normally slender frame, but Rahn was hopeful. There was no sign of the leukemia in his blood. It seemed that the worst was behind him.
The Blue Death Page 18