-7- 
Inconsistency in institutional interpretations of the guidelines seems 
inevitable. The director of ORDA recently indicated that as many as 15% 
of locally approved MUAs have contained substantive errors, including 
errors of classification. Dr. Fredrickson implies that this is a price 
that should be paid for institutional responsibility. 
I do not suggest that authority to approve even minor changes in experi- 
ments should remain at the Federal level. I do recommend, however, that 
prior NIH approval be required wherever the appropriate containment levels 
nay be subject to different interpretations. If no judgment is required to 
assign containment levels to particular experiments, institutional approval 
is appropriate. However, NIH should review such actions promptly and 
require correction of any errors within a specified period of time. The 
same procedures should be applied to both NIH-funded and pri vately - supported 
research projects subject to the guidelines. 
Surprisingly, the Federal Register documents contain no indication that 
there have been two thoroughly investigated violations of the guidelines. 
One involved the use of a vector prior to its certification by NIH and 
continuation of the experiment for some time after discovery of the error. 
In the other case, research proceeded for several months without an 
approved MUA. In both instances, the institutional committees as well as 
individual investigators were at fault. The duties of IBCs in section IV 
are described in much too general terms. More detailed expectations must be 
[A-425] 
