Pharmaceutical company InterMune performed a study on an FDA-approved drug, Actimmune, to see if it could be used to treat chronic lung disease. The study results were mixed, but if one looked at a subgroup of patients with mild to moderate lung disease, the study showed a statistically significant benefit. CEO Scott Harkonen trumpeted these results in a press release—and faced criminal prosecution as a consequence. A jury agreed that the press release was fraudulent, and convicted him in 2009, though acquitted him on the government's off-label marketing charge. (One suspects this is indicative of a compromise verdict, but the press coverage doesn't say if this was raised at any point.)
On appeal, in an unpublished opinion, the Ninth Circuit refused to hold the press release protected by the First Amendment, pointing out that the jury found the communication false. Harkonen's attorney complains "Allowing the government to criminally prosecute and convict a speaker for expressing a scientific opinion with which the government disagreed represents a real sea change in the law." [Reuters]
The result confirms my fears about the narrow application of Caronia. Who is going to want to risk their freedom on whether a lay jury correctly assesses whether a statistical analysis is intellectually honest? (Most lawyers and judges aren't equipped to evaluate legitimate controversies over statistical analysis.) The effect will be disastrously chilling, and disincentivize the dissemination of scientific knowledge.