When Scientists Assume

Here's a nice illustration of the importance of lay knowledge in risk management:

The FDA detected melamine and its byproduct cyanuric acid separately in four of 89 containers of infant formula tested in the fall, but never at the same time. A can of milk-based liquid Nestle Good Start Supreme Infant Formula with Iron contained traces of melamine while three different cans of Mead Johnson's Enfamil LIPIL with Iron had traces of cyanuric acid.

The FDA says studies show potentially dangerous health effects from the industrial chemicals only when both are present. The lack of dual contamination is key, say agency officials, and thus there have been no recalls of the tainted formula.

In a letter Friday, consumer advocates told FDA commissioner Andrew C. von Eschenbach and U.S. Department of Health and Human Services Secretary-Designate Tom Daschle that they were concerned the FDA was assuming parents would never feed their babies more than one type of formula. They said they had heard from a concerned mother who routinely fed her baby two different formulas because "one caused constipation, and one caused loose bowels, but together the baby's digestion seemed just right."

It's easy to think -- especially in the case of industrial chemicals like melamine -- that we should rely strictly on science to tell us how dangerous something is. The alternative is typically conceptualized as laypeople doing their own risk assessments, which run the danger of relying heavily on anecdotal information or unfounded assumptions (which is not at all to say that lay epidemiology is necessarily invalid).

Harder to dismiss is lay knowledge that exposes unwarranted assumptions being made by scientific risk assessors. In order to make an assessment of the risk of an activity or product, we must know both the mechanisms by which it causes harms (the chemistry of melamine and the physiology of its ingestion, in this case) as well as the social practices by which the activity is carried out and safety measures may be implemented. These latter areas are not ones that scientists can claim any particular expertise in. Too often, they complete their risk analyses using unfounded assumptions about the social side -- typically by assuming that products are used as intended by the manufacturer. Then scientists' expertise in the first half of the risk assessment is taken to justify the whole package.