UNIVERSITY OF WASHINGTON 
SEATTLE, WASHINGTON 98195 
Program in Social Management of Technology 
May 8, 1981 
Dr. Donald Fredrickson 
Director 
National Institutes of Health 
Bethesda, MD 20205 
Dear Dr. Fredrickson: 
Proposals to weaken the application of the Recombinant DNA Guidelines are 
currently before you, and I am writing to ask that you not promulgate any of 
these changes until adequate risk assessment studies have been performed. 
I am referring to the RAC actions of April 23-24, 1981 to exempt experi- 
ments in the III-O category and to end the prohibition on experiments 
with genes coding fortoxins. In addition, I have some comments on the 
proposal by Drs. Baltimore and Campbell to convert the Guidelines into 
a recommended code of practice. 
As I wrote to Dr. John Nutter of NIAID on December 10, 1981, the NIH 
"Risk Assessment Program" is seriously flawed in two major respects. 
First, it is only concerned with physical/environmental/health risks 
and these are not the only ones presented by r-DNA work. In addition to 
ethical and political risks and considerations, recent violations of the 
Guidelines should make us sensitive to the risks of intentional misconduct, 
accidents, human fallibility, carelessness, arrogance. I suspect that the 
DNA "incidents" in the future (and surely these will occur) will be based 
on such human foibles at least as often as on the failure of technological 
equipment, etc. And yet, no work is being done to assess such risks 
and mitigate their adverse aspects. The recent OTA report on Applied 
Genetics was similarly facile in this regard. 
Second, most of the NIH Risk program uses "evidence" from experiments not 
designed to evaluate risk at all . A scientific experiment cannot generally 
be said to prove a proposition which is not part of the initial hypothesis 
or, more importantly, which does not contribute to the research design. 
Can you imagine the scandal if NIH peer review sections approved funds for 
research where the experiments were designed and conducted for other 
objectives? Researchers do not make the necessary observations, perform 
relevant alterations to experimental conditions, collect requisite data 
or analyze the data from a risk perspective when their objective in 
carrying out an experiment is not risk analysis. 
314 Guggenheim, FS—13 I Telephone: (206) 543—7029 
[ 594 ] 
