Executive Summary

The result? Billions of dollars are being spent to eliminate trivial risks to health and safety based largely on speculative fears that man-made sources of chemicals and radiation are important causes of human cancer. If this same money were spent effectively, it could save 60,000 lives each year and thereby add 600,000 life-years to the life expectancy of the American people. (A "life-year" saved is a statistical measure of how much a lifesaving program increases the life span of a target population.)

For example:

• Spending $100 million per year on control of benzene emissions at rubber tire manufacturing plants might save one life-year over a 200-year period (i.e., $20,000 billion per life-year saved). The same $100 million, if invested in automobile airbag technology, is expected to save 2,000 life-years every year (or $50,000 per life-year saved)!
• Spending $100 million per year on control of routine low-level releases of radiation from nuclear power plants might save one life-year each year. But the same amount of investment in cervical cancer screening and treatment is expected to save 2,000 life-years every year.

• The median medical program costs $19,000 per year of life saved; the median injury prevention program costs $48,000 per year of life saved; and the median toxin control program costs $2.8 million per year of life saved.
• Put another way, the median toxin control program costs 58 times more than the median injury control program (per year of life saved), and 146 times more than the median medical program.

This perverse pattern of investment amounts to "statistical murder" of American citizens. Policy makers need to ask harder questions about whether our public health and environmental protection dollars are well spent. Legislators should pass broad-based legislation requiring use of risk analysis and cost-benefit analysis in governmental decisions. The president and Congress should reexamine annual appropriations to public health and environmental agencies to determine how reallocations of dollars could offer more health protection at no greater cost to the taxpayer or private sector. And communities faced with residual risks from exposures to toxins should be given greater flexibility to reduce risks through cost-effective measures.

Introduction: Health Care and Toxin Control1

Meanwhile, toxin control - the regulation of toxic substances such as chemicals and radiation - consumes about $200 billion per year, yet escapes the scrutiny health care spending receives. The public debate about environmental spending in general and toxin control in particular is far less sophisticated. Cost restrictions are fewer and public understanding of benefit issues is lower. Some influential advocacy groups demand that government protect the public against all environmental toxins regardless of how small the risks or how great the costs. Even some industry groups support legislation that would forbid regulators from weighing the benefits and costs of alternative regulatory actions. This "zero risk" or "negligible risk" perspective informs the U.S. approach to toxin control under federal laws including the Clean Air Act, the Clean Water Act, the Safe Drinking Water Act, the food additive provisions of the Delaney Clause to the Federal Food, Drug and Cosmetic Act,2 and the statutes governing cleanup of hazardous wastes (Superfund and the Resource Conservation and Recovery Act).3

"Some groups demand that the government protect the public against all environmental toxins regardless of how small the risks or how great the costs."

The public is only beginning to grasp the costs of such an idealistic approach to environmental policy. The annual rate of increase in toxin control spending recently outstripped the annual rate of increase in health care spending.4 Policymakers need to ask harder questions about whether our environmental protection dollars are being well spent.5 If the nation does not begin applying risk analysis principles to toxin regulation, we may face a fiscal crisis in environmental policy as severe as the one we face in health care policy. The failure to compare the costs of toxin control rules to rules on health care and injury prevention and to allocate resources based on those comparisons is resulting in "statistical murder." In the United States, an additional 60,000 lives could be saved each year if we applied the same cost-effectiveness standards to all lifesaving programs and reallocated monies accordingly.

This study compares the cost and effectiveness of selected toxin control measures to the cost and effectiveness of selected medical procedures and injury prevention programs. The figures used are derived from the "Lifesaving Database," a computerized information system created and maintained by the Harvard Center for Risk Analysis (HCRA) under a grant from the National Science Foundation.6 The study discusses the uncertainties in cost-effectiveness figures as well as quantitative and qualitative factors that complicate straightforward comparisons. It concludes with some steps that decision makers can take to reallocate resources from wasteful to worthwhile programs.

"Cost Per Life-Year Saved":

A Yardstick for Comparison

"Life-years saved" represents the impact of premature death on an average American's life span. For example, those who die of cancer at age 65 may lose 15 or so years of life expectancy. Consumers have limited opportunities to directly purchase "life years" in the marketplace. However, based on answers to survey questions and inferences about routine on-the-job safety decisions, health economists have estimated the value of an average life-year, based on our willingness to pay, at somewhere between $10,000 and $500,000.8 An underground coal miner or an ironworker on high-rise buildings, for example, would command higher wages than a clerk because of the higher risk involved in the job. The 50-fold range reflects variation in citizen preferences as well as genuine uncertainty about how much people care about life expectancy.

"Some toxin control regulations impose astronomical costs for a year of life saved."

There is little evidence that cost per life-year saved is a significant consideration in toxin control regulations, some of which impose astronomical costs relative to their benefits. For example:

• l In a recent survey of 10 Environmental Protection Agency (EPA) rules aimed at curbing air emissions of benzene, a chemical known to cause leukemia at high doses, the estimated cost per year of life saved ranges from $200,000 to $50 million.9
• l A proposed EPA standard for chloroform emissions at some pulp mills costs more than $99 billion per year of life saved.10

There is both more reason and more opportunity than ever to subject toxin controls to some measurement of cost-effectiveness.

The Cost-Effectiveness Framework

• Effectiveness is defined as the increase in life expectancy (i.e., years of life saved) resulting from a program.
• Cost is defined as the material and labor resources, measured in dollars, necessary to implement the program minus any resulting savings in resources (taking into account resource impacts in both the public and private sectors of the economy).12
• The cost-effectiveness ratio is the estimated net program cost divided by the estimated number of life-years saved (in comparison to a well-specified program alternative).

"The Lifesaving Database makes it possible to compare the cost-effectiveness of programs."

Figure I is a frequency distribution of the cost-effectiveness ratios for all 587 lifesaving programs. The median program costs about $42,000 per year of life saved, although the ratios range over more than 10 orders of magnitude. Thus, the range is from approximately zero cost up to $100 billion per life-year saved. Most of the programs cost between $10,000 and $1 million per year of life saved.

"Most of the programs cost between $10,000 and $1 million per life-year saved - but the cost ranges up to $100 billion."

Favorable Cost-Effectiveness. About 10 percent of the programs have net costs equal to or less than zero, which means that the program saves resources equal to or greater than the resources it consumes. For example [see Table I]:

• Immunizing children against mumps, measles and rubella is cost saving because the estimated reductions in subsequent treatment costs more than offset the costs of immunization.
• Requiring stricter flammability standards for children's sleepwear size 0-6x is another program that reduces estimated costs more than the costs of implementation.

• The median medical program costs $19,000 per year of life saved.
• The median injury prevention program costs $48,000 per year of life saved.
• The median toxin control program costs $2.782 million per year of life saved.
• Put another way, the median toxin control program costs 58 times more per year of life saved than the median injury prevention program and 146 times more than the median medical program.

Information on the medians does not tell the whole story. Figure II compares the frequency distributions for medical and toxin control programs, illustrating that the range of cost-effectiveness ratios is large for both. Some toxin control programs save more resources than they cost. For example, the accelerated phase-out of lead in gasoline has been estimated to save more in automobile maintenance expenses than it costs in extra refining investments.17 But note also that the frequency distribution for the toxin control programs is shifted to the right in Figure II, suggesting that the average toxin control program costs much more per life-year saved than the average medical program. As Figure III shows, the frequency distributions of injury prevention and medicine are similar.

"Some toxin control programs save more resources than they cost, but the average program costs much more than the average medical or injury prevention program."

How To Save 60,000 Lives

• If resources were reallocated to more cost-effective programs, an additional 60,000 lives (or about 600,000 years of life) could be saved each year in the United States at no increased cost to the public or private sectors.
• Looked at another way, an efficient reallocation of resources could save $31 billion per year without reducing the number of lives (and life-years) saved.
• To demonstrate why the allocation of resources is so important:
• We spend $115.6 million per year on benzene emission control during waste operations to save five life-years.
• Spending the same amount on collapsible steering columns in cars saves 1,684 life-years.

Table I shows 10 programs that save both money and life-years. By contrast, Table II shows the 10 most expensive programs relative to each year of life added. As the tables show, the EPA standard for chloroform emissions at 48 pulp mills imposes over $99 billion in costs for each year of life added, while the standard at 17 low-cost pulp mills actually saves resources. This information, while neither definitive nor exhaustive, suggests that decision makers should scrutinize how lifesaving resources are being allocated.

Is It Fair To Compare Toxin Control to Other Health Programs?

Objection No. 1: The lifesaving effectiveness estimates for toxin control programs are less certain than those for medical services.

• Medical effectiveness estimates tend to be based on randomized trials of treatments or direct epidemiological observations.
• Toxin control effectiveness estimates are often based on results extrapolated to humans from high-dose rodent tests of chemicals.

Despite this uncertainty, there are no indications that the risk-reduction predictions for medicine are systematically more optimistic than those for toxin control programs. Rather, there are indications that the baseline estimates of risk from toxin exposures are often more pessimistic than those for medical technologies. For example:

"There are no indications that risk-reduction predictions for medicine are more optimistic than for toxin control."

• The baseline estimates of cancer risk from exposure to toxins are based on the assumption that any exposure to a carcinogen, however small, increases one's cancer risk.
• Recent scientific evidence on chemicals such as chloroform, formaldehyde and benzene suggests that this assumption is incorrect and that low levels of exposure are associated with little or no incremental risk of cancer.19

Objection No. 2: The focus on saving lives ignores health effects that are serious but not fatal.

In general, programs that save lives may also reduce nonfatal cases of disease or trauma that range in severity from several days of discomfort per year (e.g., from mild bronchitis) to chronic pain and discomfort (e.g., chest pain from angina) to extended periods of disability and/or illness requiring hospitalization. The next generation of studies likely will include these "morbidity effects" in a new measure of effectiveness called "quality-adjusted life years" (QALYs) saved. This measure, which is increasingly used in medicine, combines information on life expectancy and on quality of life.21

Objection No. 3: The focus on human lifesaving ignores the possible impacts of toxins on nonhuman species.

Objection No. 4: The cost estimates for toxin control programs are more uncertain than those for medical programs.

"No systematic evidence supports the contention that cost estimates are more uncertain for medicine than for toxin control."

Objection No. 5: The risks from exposure to toxins may be ethically and/or psychologically more compelling because these risks are unfamiliar, invisible, unfair, frightening and imposed on citizens without their knowledge or consent.

Practical Steps for Advocates and Policymakers

The answer is that no matter how strong the case for public health and environmental protection may be, the voting public will limit the resources available for risk reduction. They will do so precisely because they also care about other things - public safety, crime prevention, housing quality, education quality, access to recreational opportunities and so forth. Further, what some regard as "waste" in the defense budget is regarded by others as essential to preparedness.

In the end, some rationing of resources is inevitable.

If public health professionals attempt to protect inefficient investments, they undermine the long-term credibility of public health policy. Most inefficiencies are exposed eventually, and their exposure tends to reduce public confidence - and the budgets the public is willing to support.27

The failure to terminate inefficient public health policies also leads to "statistical murder," as mentioned above. Skeptics will challenge this claim. If the nation decides against a tightening of benzene emission standards at oil refineries, they will argue, this will not result in more prenatal care or fewer cases of AIDS. They are right, unless policymakers assure the desired transfers of resources. There are three concrete reforms policymakers can begin to make now.

Reform No. 1: Reallocate tax dollars to those programs that pay the greatest health returns.

Reform No. 2: Void and avoid mandates that would force the private sector to spend money where the public sector should not.

"If an expenditure is inappropriate for the federal government, it is no more appropriate for the private sector."

Reform No. 3: Give local communities flexibility.

Conclusion

NOTE: Nothing written here should be construed as necessarily reflecting the views of the National Center for Policy Analysis or as an attempt to aid or hinder the passage of any bill before Congress.