1. The jury in the Barnett case was asked both if Merck should be held strictly liable for failure to warn and causing Barnett's heart attack, and if Merck was negligent in failing to warn and causing Barnett's heart attack. Interestingly, the jury answered the first question "No," and the second question, "Yes," which is logically inconsistent. While it's possible to not be negligent, yet held strictly liable, it's impossible to both be negligent but not strictly liable.
Merck indicated in a conference call yesterday that they would challenge the verdict on those grounds, but, if so, they're likely to be disappointed: the remedy for an inconsistent verdict is to ask the jury to return to deliberations to reconcile the inconsistency before they're dismissed. Otherwise, the issue is usually considered waived. I haven't researched the Fifth Circuit's jurisprudence on this question, so it's possible they come to a different answer, but the general rule here is that the inconsistency doesn't negate a verdict unless it's challenged at the time. (Because I don't have the trial record in front of me, it's possible that Merck did challenge the inconsistency at the time and Judge Fallon incorrectly refused to order new deliberations, but that wasn't the sense I got from the press conference.)
2. Barnett had $150,000 in medical expenses. So almost the entirety of the $50 million compensatory award is non-economic damages.
3. Barnett's attorney argued to the jury that Merck should face punitive damages because they dared to try to defend themselves in court instead of settling the cases. Fallon upheld an objection to this tack, and agreed to consider striking punitive damages because of the improper argument.
4. Mark Lanier's spin of the Barnett and Humeston decisions is on Peter Lattman's blog, and there's an interesting discussion in the comments where I've contributed.