EPA Policy Choices : ‘Managing’ Risks: Sense and Science

The matter at hand involved uranium mill tailings, tens of thousands of tons dumped in towering piles throughout the West, leaking radiation.

Cover them with dirt, the Environmental Protection Agency decided. Seal off the radioactivity.

But then came the question of how much dirt.

The lingo at regulatory agencies such as the EPA includes the term “best available technology.” Well, reasoned Milton Russell, then an EPA assistant administrator, the best available technology here is to pile on more dirt, and more dirt and more dirt.

Drawing the Line

But it was impossible to pile on an infinite amount of dirt, or reduce the radiation risk to zero. So where, Russell wondered, would be the stopping point?

In this fashion a seemingly technical question about dirt in time became a fiercely argued issue loaded with the imponderables of moral philosophy.

How to make regulatory decisions affecting human welfare represents one of the most controversial and difficult tasks in modern times. The decisions carry with them huge significance. They are relied upon to protect and save lives. They determine the expenditure of billions of dollars. They affect elections. They fill the courts to overflow. They generate and provide grist for powerful environmental movements and corporate lobbies. But those making the decisions generally acknowledge that they work largely in ignorance and ambiguity, relying on subjective judgments.

This is a look inside the process that reaches such decisions, as seen by one who participated. For three years and eight months, until he stepped down last March, Milt Russell, 53, served as assistant administrator for policy, planning and evaluation at the EPA. Put more simply, he was what is now called a “risk manager.”

Bound to Be Unpopular

Russell is a Republican appointee and an economist who talks like one, describing pollution as an “externality of production.” So environmentalists more often than not have found themselves at odds with him. Yet it is fair to say that risk managers in any form would not be popular figures.

A risk manager, after all, is in the business of recognizing trade-offs. His underlying premise is that not every good thing can be done. He balances human lives against economic cost, among other considerations, performing a sort of environmental triage. He traffics in imponderables about the value of human life and the morality of knowingly imposing involuntary risks on people.

He also presides at a shotgun marriage of science and law, something that fails to satisfy either the scientists or those who advocate, draft and administer environmental law. Many laws presume the possibility of zero risk and assign liability where that is not reached. But science’s increasingly sophisticated ability to detect and finely measure dangers renders the world a mine field of risks, many unavoidable. Science cannot provide the certainty of protection demanded by the statutes.

“You can’t wish some problems away,” Russell said one day recently, soon after leaving the EPA. “Every problem doesn’t have a solution and the solutions to those problems which do have solutions are not necessarily nice. . . . There is something about the American character that seems to assume there is a good solution to every problem, and that just ain’t true. I mean, there are just some problems for which there are no easy, much less right, much less perfect, solutions.”

Risk managers of Russell’s sort did not exist a decade ago.

Looking back now, there are those who trace their emergence to a moment in 1972 when radioactive tracer studies for the first time found residues of the carcinogenic growth promoter DES in livestock. The traces, of course, had been there since the use of DES began. They just hadn’t been noticed until the detection methods grew sufficiently sensitive.

Crisis for the FDA

This discovery presented something of a crisis to the Food and Drug Administration.

The agency, then operating on the presumption that no level of a carcinogen could be regarded as safe, had to ban DES immediately. The staggering ramifications were immediately apparent to the FDA’s general counsel then, Peter Hutt.

If DES could be detected in livestock at low levels, certainly traces of other carcinogenic drugs and feed ingredients could also be found. The FDA, it seemed clear to Hutt, faced the prospect of banning all carcinogenic substances from animal feeds and drugs. That would sweep bare most supermarket shelves, a prospect that Hutt considered enormously impractical, if not impossible.

A question occurred to the general counsel: What the hell are we going to do?

It was then, just days after the FDA withdrew approval of DES, that Dr. Richard Lehman of the FDA Bureau of Veterinary Medicine walked into Hutt’s office one afternoon.

Lehman had something in mind that he thought might be helpful in handling Hutt’s problem.

“Quantitative risk assessment,” Lehman said.

Hazard Under Study

The idea of quantifying human risk from carcinogenic substances, on the basis of mathematical extrapolation from animal feeding studies, had been examined by scientists for more than 25 years. But it had remained an inquiry for scholars, not a tool for regulators.

The practical use of risk assessment seemed quite apparent to Hutt and others. Since they were now going to find trace elements of carcinogens throughout the food chain, they needed to draw differences between significant and insignificant levels, if they did not want to end up banning the whole food chain.

In the July, 1973, Federal Register, the FDA proposed regulation that for the first time would quantify human risk from a carcinogen and establish a “safe” level of exposure. Since then the risk assessment business, responding to the needs of the times, has become a booming industry.

Over the years, scientists refined techniques and wrote mountains of research papers. They were hired both by regulatory agencies and private industry. Eventually, a good number opened their own risk-assessment consulting firms, some furnished in chrome and glass with high-tech names such as “Environ.”

There were, though, always certain limitations connected with the business of risk assessment.

Imprecise Calculations

The calculations produced by the process, most agreed, were uncertain at best and very possibly meaningless. Even those doing them allowed that they were transcending the bounds of traditional science, making judgments and assumptions to reach a number.

The leap from low-dosage animal exposures and other sorts of data to human risk estimates is made via mathematical models that depend on a host of inferences and assumptions piled one atop another as the calculations continue. Mistaken judgments can get magnified exponentially.

Industrialists, unions and environmentalists, of course, all have different opinions of this process. The calculations are variously lambasted as wildly overestimating risks, as failing to measure most risks, and as a mathematical sham designed by corporate interests to dilute regulatory efforts.

It is fair to say that the assessments lack precision. The National Academy of Sciences, for example, has cited 10 statistical models that yield strikingly different predictions about saccharin and bladder cancer. The range stretches all the way from 0.22 to 1,144,000 lifetime cases of bladder cancer resulting from consumption of one bottle of diet soft drink per day by 220 million persons over a lifetime.

Joseph V. Rodricks, a toxicologist by training, spent 15 years at the FDA, the final two as chief science adviser to the FDA commissioner. Now head of the consulting firm Environ, he has conducted dozens of risk assessments for private clients and government agencies. He is a founder of the Society for Risk Analysis and the author of more than 70 scientific papers in the field.

Science Lags Behind

He said: “We can’t pretend to get scientific results. Science is 50 years behind what the public wants us to provide. Someone has to go beyond the data. Someone has to answer the question of where to draw the line. Pure science would say ‘we don’t know if saccharin causes bladder cancer,’ but someone has to fix on an answer. The whole regulatory operation depends upon it.”

So emerged the risk manager.

Offering the EPA position to Russell in mid-1983, Al Alm, then the deputy administrator, had explained the job: when decisions loomed, he and William D. Ruckelshaus, then the agency’s administrator, did not have the time to ask whether the analyses were right, whether the assumptions were checked. They wanted never to have to worry. Russell, by his own description, was more an analyst than an ideologue, experienced in government after a stint in the mid-70s as senior staff economist for the Council of Economic Advisors. Alm thought Russell could make sure that they would never have to worry.

It was Russell’s notion that his new job could be executed in an organized and sensible process. He was not naive--he understood that the final decisions would require judgments--but he believed they could be reached through what he appreciatively called “a rational framework.”

He set up a formal “development plan.” He set up an “option selection process.” He set up meetings to consider “cross-media implications.”

System of Assessments

After six months of this, Russell felt a sense of accomplishment. A rational system was now in place.

Beginning with the risk assessment figures, he would ask certain questions: How sure is the scientific evidence? What is the degree and distribution of risk? What is the investment necessary to reduce it? Does this investment make sense, given all the other demands on our economy? What kinds of social and personal disruptions will regulation bring?

These seemed to him impeccably logical questions.

Others disagreed.

At public gatherings, his critics would rise, stabbing the air with accusing fingers. How can you stand there, they would ask, knowing there is a risk and not stopping it? How could you decide that some people will die of cancer?

Thus began the education of a risk manager.

One of his first lessons came three years ago, in the case of the uranium mill tailings.

Uranium milling produces powdery, sand-like tailings that release radon, a gas that is the second deadliest cause of lung cancer (after cigarettes). Before this danger was fully understood, the sandy tailings were used as fill dirt and building materials in house construction, particularly in the Rocky Mountain states. The tailings were also dumped at rural sites scattered throughout the West.

Congress passed a law in 1978 requiring the government to find and eliminate this contamination. At thousands of houses, many centered around Salt Lake City and Grand Junction, Colo., basements, gardens, patios, garages and foundations were torn up and replaced.

It was less clear what should be done with the tailings that simply remained dumped in soaring piles. Upon reflection, the EPA fixed on the dirt idea. It was then Russell faced the question of how much dirt.

There could not be a more appropriate example of the risk manager’s dilemma.

The tailing piles left untreated and uncovered, the EPA estimated, would probably produce 600 extra cancer deaths in a century. This is the rough equivalent of six deaths a year, although mathematical purists would say that such extrapolations from lifetime estimates are problematic.

If the EPA spent $500 million, they could pile enough dirt on the tailings to reduce the risk by 95%, from 600 deaths a century to 30, from six deaths a year to one death every 3 1/2 years.

There was an alternative to this plan. By roughly doubling the amount of dirt, they could reduce the risk further, by 99.5%. Instead of one death every 3 1/2 years, they would have one extra death every 30 years.

Costly Extra Fill

The extra dirt, though, would add another quarter of a billion dollars to the bill.

There were plenty who thought the EPA should take that second option. Some were members of environmental groups. Some were inside the EPA. Some were on Russell’s staff.

There is no right answer here, Russell thought. There are always people who have different values.

Russell sought to put the six deaths in perspective. Roughly 125,000 people die of lung cancer each year in the United States, the vast majority due to cigarette smoking.

Still, he knew, no one had asked for this radon risk. Cigarettes are far more dangerous, but people don’t inhale uranium tailings voluntarily.

Although the total incidence of cancer from the tailings was relatively low, the risk was considerably higher to the tiny population living close by the piles. How, Russell wondered, to balance large risks to very small populations against the general public interest?

Russell knew also that these tailings were the product of milling uranium for the nuclear power industry. There was anger and resentment in the community about that. These were serious people, not kooks. You could say, somebody made this mess. The power companies, they ought to pay.

Unconfusing the Issue

And yet, Russell also thought: whether you should build more nuclear power plants is a very different issue from what to do now with these tailings. You could close every nuclear power plant and every bomb plant in the country tomorrow, and the tailings would still be there and you would still have to decide what to do with them.

That extra quarter of a billion dollars needed to get the risk down from 95% to 99 1/2%--in a world of limited dollars, is that the best place to spend it for improved health? You could buy a hell of a lot of kidney machines and ambulance cardiac-care units for that. If you’re worrying about lung cancer, why not spend the quarter-billion to buy off tobacco farmers, or for an anti-smoking ad campaign, or on programs for early detection of lung cancer?

There was something else, something far less esoteric, quite distanced from the economic textbooks at the University of Oklahoma, where Russell studied.

Russell grew up on a farm in Texas. He knew about heavy equipment, knew how dangerous that stuff was to operate on the narrow dirt roads out there in the West. They would be moving millions of tons of dirt over long distances, stirring up dust, clouding visibility, filling lungs.

Comparisons of Risks

You cannot put a price on those three or six lives lost to radon, people told him. But those lives to him were just statistical creations, dubious extrapolations on a page. Russell was thinking of real guys driving real bulldozers with real families.

You could not factor that into the computer, Russell thought. But it’s there.

The pressure came from all sides. Congress had passed an amendment threatening to hand the matter over to the Nuclear Regulatory Commission if the EPA did not act. A congressman, chairman of the House Armed Services subcommittee on procurement and nuclear systems, had lashed out at the agency for setting standards that were too tough, considering that the dangers of radon were unproved. Then, when the EPA moved to relax those standards, David Berick at the Environmental Policy Center said he found the move “disturbing.”

The EPA eventually decided to spend half a billion dollars and pile on enough dirt to reduce the estimated risk by 95%. It did not make sense, the agency decided, to spent a quarter-billion dollars more to get the risk down by 99 1/2%.

Russell had come to regard part of his job as applying the “laugh test” to decisions. After all the technical analysis, does the decision hold up under common sense?

Could Pass Laugh Test

This one could pass the laugh test, Russell figured.

There were lawsuits. Ruckelshaus took the brunt of the vilification, of course, but Russell caught his share at small community meetings. He was uncaring, they said, unconcerned, a monster. He was letting go unconscionable risks that could be remedied.

Well, Russell thought: that’s just one of those kinds of decisions you end up making.

He also thought: making a decision with your eyes open and in front of everybody is going to be very difficult. It’s easy to make decisions. It’s difficult to make good decisions and to make them in the light of day.

Ruckelshaus soon came to Russell with a suggestion: Go out there and do something about getting reasoned decision-making in the American public. Go out there and educate.

Russell took to the road, delivering speeches. Russell knew where he stood and felt willing to spell it out.

Defining the Acceptable

Acceptable risk is an uncomfortable term, he would say at the podiums. The popular notion is, no risk is acceptable. But public officials who talk about delivering zero risk perpetuate a comforting myth. On the other hand, to question that myth opens one to charges of being callous and uncaring. Not an appealing prospect. Little wonder “acceptable risk” is a topic public officials most often reserve for quiet conversations with intimates.

“To risk managers and those comfortable with modern technology, pollution is an externality of production,” he said in one speech. “Its reduction is a goal of public policy to be pursued vigorously, to the point that it maximizes the total public interest. To others, reflecting more traditional, deeply held values, it is a violation of personal rights; hence its production is an evil and its reduction a matter of moral principle. For them, to speak of an optimal level of pollution is like being told that each community should have an appropriate number of child molesters. . . .

“When you pierce through the fog of posturing and politics and special interests, the profound dichotomy between pollution risk as a factor of production and pollution risk as an evil lies at the root of much environmental debate in the country.”

Yet Russell understood that he was laboring amid ambivalence. Up on the podium, he also said: “The idea of some risk for the greater good conflicts with the position that each individual is infinitely precious. This is one of the eternal conundrums of man in society. I doubt that it will be resolved by the EPA.”

A Different Approach

The EPA tried a different approach in Tacoma.

The old copper smelter there had sat on the shores of Puget Sound, smack in the middle of the city, for about 90 years. Arsenic only recently had been discovered to be a volatile human carcinogen. EPA analysts figured that emissions from the plant, unregulated, were causing four additional lung cancer deaths each year.

To eliminate the arsenic emissions would require controls so stringent that the plant would just close down, unable to operate profitably. That would mean the end of 570 jobs, a $30-million payroll. You could instead cut emissions almost in half and keep the plant open, but the smelter would still cause one extra lung cancer death beyond the norm each year.

Economic conditions in Tacoma were fairly distressed. Russell thought: You take away those jobs, it’s fairly well documented you end up with increases in alcoholism, child abuse, spousal abuse. You also end up with fear. That itself is a public health concern.

And yet: The wind blew from the plant toward an affluent suburb where people had no need for smelter jobs.

Asking the Public

The EPA opted for an experiment. The agency would communicate all the options and alternatives and uncertainties to the community. The agency, of course, would make the final decision, but first it would let the people indicate where they would make the trade-offs.

So Russell found himself at a public hearing.

First, a fellow in his 70s rose to speak. He had been working at the smelter for 40 years, he said. He had consumed so much arsenic in his day that his sweat was green. How can you say this arsenic is hurting anybody? He had lived with the stuff, he’d eaten it, he’d drunk it. What are you talking about?

Then a young, frail woman rose, wisps of hair falling on pale skin, eyes full of tears. Her child had died of leukemia. That stuff in the air had killed her child. What was the government going to do about that?

It occurred to Russell that he was facing people different from himself. He was a man of rational analysis listening to people of faith.

The agency did not want the matter cast as a jobs-versus-health issue in Tacoma, but that is how it seemed.

Buttons Tell Sentiment

The crowd showed up one day at a hearing wearing newly pressed buttons.

“Both,” they said.

If we had succeeded in explaining all this, Russell thought, the buttons would have said: “Some.”

The EPA never did have to decide this one. Before the agency could act, the smelter owners closed the plant. The copper market had gone so bad it wasn’t profitable under any circumstances.

The EPA, nonetheless, got blasted in Tacoma. Why is the EPA asking us to make these types of decisions, a good number in Tacoma asked. They’re the experts. Why can’t they make the decision? This is confusing.

This is confusing to us, too, Russell would respond.

They were no longer the people wearing the white hats. There were no objective right-and-wrong, yes-and-no answers. They had to make value judgments and expose their own uncertainty. They could no longer regard themselves as experts.

That, Russell observed, is a dangerous thing for your psyche.

Campaign on Gasoline

In the end, it was another issue--lead in gasoline--that provided Russell his greatest sense of both triumph and frustration while at the EPA.

Lead-free gasoline had been introduced in the 1970s to accommodate newer cars with catalytic converters. The lead levels in regular gasoline had been reduced, but you could still find gasoline with lead at the pumps in the early 1980s.

Some used it for older cars without catalytic converters. Some used it illegally in newer cars with converters. Leaded gasoline, after all, had more octane and was cheaper.

Lead in gasoline gets emitted from tailpipes, polluting the air, water, street surfaces. The streets had a particular import. Kids pick up dust from the street, stick their hands in their mouths, end up ingesting the stuff.

Russell found that he could correlate the kids’ blood lead levels with the increased driving during summer months, the relationship was so direct. He could also draw lines between the blood levels and horrible maladies--brain dysfunctions, behavioral disorders, biophysical problems.

This one was not ambiguous to Russell, for once. His own department had initiated and pushed the research. When they toted up the pure cost-benefit figures, the benefits came out roughly three times as high as cost. That was without even trying to put a price on things like pain and suffering and decline in children’s IQs.

The decision was easy. They laid it out before Ruckelshaus, took just an hour and a half. Let’s roll, the administrator had said.

There were no denunciations or lawsuits this time. But there also was no applause. What Russell considered the most significant protective action during his tenure, one that salvaged hundreds of thousands of lives, was largely met by public indifference.

Russell felt he understood.

Getting the lead out was going to raise the price of gasoline at the pumps and take away some octane, for an issue people didn’t know was serious.

“It’s not baby seals,” he said later.

He also said: “And who’s the enemy? There was no place that anybody could trumpet victory or describe evil. It was a classic case of a social benefit, social good, for which there were no villains and for which there were no heroes.”

The risk manager’s frustration was understandable. Public perceptions do differ markedly from what science says are the biggest risks.

Hazardous-waste sites trigger intense reactions, but scientists say that household radon, air pollution and destruction of the earth’s ozone layer pose far greater threats.

Perceptions of Risk

The public, when polled about risks, tends to list such items as nuclear power and food preservatives near the top, while statistical analysts would put those items far down on a list that is headed by smoking, alcohol, automobiles and handguns.

States sent state troopers to clear supermarket shelves of foods sprayed with the pesticide EDB, a risk measured in parts per million, while the everyday world is full of comparable risks. Drinking a half-liter of wine increases the chance of death over a lifetime by one in a million. So does living two days in New York or Boston, traveling 10 miles by bicycle, eating 40 tablespoons of peanut butter or 100 charcoal-broiled steaks.

It could be said, though, that Russell’s rationality had come up against something more than a perception problem. It could be said he was bucking the perplexing force of human nature.

Those who study such matters report that the risks people care about have their own type of logic, a logic that transcends the data.

People are personally offended when someone is imposing a risk on them. A small risk involuntarily imposed is more objectionable than a much larger risk voluntarily assumed. One study, in fact, indicated that people are willing to accept a risk 1,000 times greater if it is of their own choosing. A person elects to ride a bicycle or smoke a cigarette. He does not elect to ingest EDB. Equity is the issue.

Familiar Risks Favored

People will also accept controllable or familiar risks--swimming, X-rays--over what is beyond their control or unfamiliar. Paul Milvey, a senior analyst for the EPA, an expert in risk calculations, with a doctorate in biophysics, relates a telling story about himself along this line.

He was in a Washington area hospital, facing the need to undergo a CAT-scan with contrast media. Because the test involved injection of an iodine-based dye, his doctor explained, there was a slight risk of death if his kidneys failed. They could instead take a nuclear magnetic resonance picture, a test that was not invasive and did not carry any risk, but this would require a trip to another hospital in the Washington area.

By all means, Milvey said, he would prefer going for that other test.

Then his wife pointed it out to him: The risk of death in a traffic accident while traveling by ambulance to the other hospital was greater than the risk of the CAT-scan.

I’d still prefer the ambulance ride, Milvey said.

In the end, he needed neither test, for the doctors decided that he was healthy. But he cannot get over his reaction.

Reflects on Lesson

“Even though I knew instantly my wife was right, I nevertheless preferred the option of the ambulance risk to the unknown danger,” he said. “It amazed me that although I’m trained to look at these things objectively, I couldn’t do it, I couldn’t convince myself otherwise.”

The message here is one that Russell, however reluctantly, has come to acknowledge. Ever the analytical economist, he sees it as part of his equation.

“It is perfectly reasonable and acceptable, the way the world is, that some kinds of risk are cared about more than other kinds of risks,” he said one day recently. “That’s a datum for the EPA, and data points are not something you argue about. Whether they are rational, irrational, good or bad, they are data points. They have to be dealt with.”

In mid-March, one week before he left the EPA, Russell traveled to Charleston, W. Va., to give a farewell address of sorts before the National Institute for Chemical Studies.

“There are myths each of us live by,” he said to that group. “Let me alert you to some of these myths.”

There is the myth that risk numbers represent reality. There is the myth that there can be zero risk. There is the myth that environmental quality comes free.

Myth of ‘Answers’

Then, he said, there is the myth often held by economists and public policy professionals such as himself.

“That is the myth,” he said, “that facts and logic--rational analysis--provides ‘answers’ when it comes to environmental protection.”

This insight had evolved gradually, by degrees. There was no blinding epiphany. But Russell during his tenure at the EPA had arrived at a conclusion that marked something of a journey from where he began.

A challenging new opportunity now beckoned at the University of Tennessee and the Oak Ridge National Laboratory in Knoxville. The pace would be less intense. He had found the EPA job fun, challenging, exciting. There had been a certain measure of satisfaction. It had been a very good 3 1/2 years, full of amazing changes, things accomplished.

But it did seem like he had been there forever, and people in his type of position do get tired. He was not naive, but he had underestimated the resistance to dealing with matters in a systematic way.

Now, he said, “Maybe it’s time for somebody else to make the decisions.”