While the regulations at 42 CFR 50.103(d)(14) are vague with
respect to what actions are required when MIS is substantiated, such as
withdrawal or correction of publications, the ORI *Model* Policy is
clearer about the route ORI considers desirable.

 "If the Deciding Official determines that the alleged misconduct
is substantiated by the findings, he or she will decide on the
appropriate actions to be taken, after consultation with the Research
Integrity Officer.  The actions may include:
 --withdrawal or correction of all pending or published abstracts
and papers emanating from the research where scientific misconduct was
found;
 --removal of the responsible person from the particular project,
letter of reprimand, special monitoring of future work, probabion,
suspension, salary reduction, or initiation of steps leading to possible
rank reduction or termination of employment;
 --restitution of funds as appropriate."

 You would want to check your institutional policy, which would
have been approved by ORI, to see if it specifically requires contacting
the journals if the MIS decision relates to the factual aspects of the
published info.  If the policy is silent, I think it would still be
prudent to correct/retract.  First and most importantly, it shows "honor
in the discipline" to correct mis-information if known.  Second, it
could bolster a defense against liability should that published
erroneous information lead to negative consequences (e.g., mis-treating
an illness).  Third, journals want to know about these situations, since
their reputations depend on the reliability of what they publish.  I
suspect they would be disturbed if not notified.  Fourth, it reinforces
to whistleblowers that the institution takes a firm stance in known MIS
cases, and lets respondents --and potential respondents-- know it is an
especially serious matter.

Celia S. Walker, Director
Regulatory Compliance, Colorado State University
Fort Collins, CO  80523-2046
970.491.1563  FAX 970.491.2293
xxxxxx@research.colostate.edu