by Jane M. Orient, MD

Pesticides on fruits and vegetables . . . asbestos in the heating ducts of our schools . . . . toxic waste disposal facilities . . . chemicals in the drinking water . . . radiation . . . side effects of vaccines . . . What should we worry about first?

Of course, it isn't just a question of worry. When health and safety are threatened, the first response is often: ``There ought to be a law.'' If there is a threat, why not just outlaw it? Reduce it to zero: ban Alar, rip out the asbestos, shut down anything that releases radiation, and so on.

The problem is that risk can never be reduced to zero. And in public health, as in medicine, sometimes the treatment is worse than the disease. The price of decreasing one risk may be to increase another. That's why it's important to know what to worry about most.

Defining an ``acceptable'' risk is bound to be controversial. Some will think the standards are too lax, and some will think they are too stringent. A general rule of thumb is that a risk of 1 in 100,000 requires action if the risk is ``involuntary.'' (Pesticide residues on food are considered an involuntary risk.) For comparison, a person has a 1 in a million chance of being killed by a natural hazard such as lightning in any given year.

If the risk is occupational (and thus ``voluntary,'' for a person can change his occupation), a level of 1 in 10,000 is ``acceptable.'' Many workers assume higher risks than that. Every year, a policeman has a 2 in 10,000 chance of being killed on the job. For miners and construction workers, the risk may be up to 6 to 7 in 10,000 per year.

In news reports about public health fears, it is usually difficult or impossible to find any numerical risk estimates.

This is sometimes because the numbers are not known. In fact, the risk may be purely hypothetical and unproved (as with the growth regulator Alar).

Often, the numbers are so small that they wouldn't sell newspapers.

One unit of measurement of the risk from chemicals is the HERP, the Human Exposure/Rodent Potency Dose, the percentage of the dose that has been found to cause cancer in rodents. The average daily intake of the banned pesticide ethylene dibromide would have given a lifetime HERP of 0.0004%, corresponding to a cancer risk of about 4 in a million. Other representative HERPs:

This is the most important question to ask when trying to determine the significance of a certain risk.

Common answers include: ``higher than average,'' or ``twice the expected,'' or ``twice the EPA standard.'' Statements like these can be used to stimulate fears of cancer. But what do they really mean?

By definition, most things are either ``higher than average'' or ``lower than average.'' Of course, a ``lower than average'' cancer rate in homes that have a ``higher than average'' radon level (as has actually been observed) doesn't prove that radon cures cancer!

``Twice the expected'' level might be evidence of a ``cancer cluster.'' But most of the time, it turns out to be due to chance. As in poker, a player with a full house might have cheated, but he might have been lucky.

The significance of a level that is twice as high as a standard depends on what the standard is. Two times a very tiny number is still very tiny.

It is much more helpful to compare one hazard with another. As shown above, eating peanut butter is more risky than eating residues of ethylene dibromide (although still not very risky). And mining is more risky than law enforcement work.

Still better is to compare one risk with the risk of the available alternatives. For example, the risk of ripping out asbestos might be far less than the risk of leaving it in place. The risk due to asbestos in brake linings might be much less than that of the available substitutes, especially if the brakes don't work as well.

This type of analysis is particularly important in medicine. While people worry about the danger of pertussis vaccine (or can't afford it due to the high cost of the manufacturer's liability insurance), twice as many children are suffering from whooping cough. People have forgotten that whooping cough once killed more children than measles, polio, scarlet fever, and many other childhood diseases combined.

Obviously, money spent for one purpose is unavailable for other uses. Each time we declare war on one danger, less can be spent fighting another. Each time a law increases the costs of production, less of a product is available and at higher cost. Alar-free apples are more expensive, to name just one example. And of course, the higher costs are a greater burden for the poor.

To make wise public policy decisions, it is essential to express costs in a way that permits comparisons. One such method is to calculate the number of dollars spent per life saved. Lawmakers can then choose. Should they vote to spend $2,500,000 to save a life by reducing the radium content of drinking water, or to spend $10,000,000 to save a life by tightening the standards for disposal of radioactive waste? (The likelihood that a person will get cancer from either of these dangers is extremely small.) Or would the money be better invested in highway construction and maintenance, which can save a life for only $20,000 (highways being so much more dangerous than low-dose radiation)? Or in passive automobile restraints, which cost about $300,000 per life saved? Or for cancer screening, which can save a life for perhaps $30,000?

Unfortunately, legislation is seldom preceded by an economic or public-health analysis. ``Protecting'' people from what they most fear gets more votes than protecting them from what is most likely to kill them.