This strange set of bedfellows comes from the just-decided patent infringement case known as Mayo Collaborative Services v Prometheus Laboratories, Inc; the patents in question are 6,335,623 and 6,680,302. The patents are fairly complex and detailed, but the infringement question seems straightforward. A good, if quick summary of what was decided today comes from Timothy Lee, writing for Ars, "Supreme Court saves medical profession from diagnostic patents".

Lee, who has written about this case for Ars before, played a small part in the Cato institute amicus brief, and is a strong advocate for the invalidity of the patents. I have not yet read the opinions in the case, but based on the summaries I believe I would have to side with Lee and the other amicus parties. To put it bluntly, this looks ridiculous on the face of it and it's not clear to me why the patent was issued in the first place, or why the lower courts upheld it.

The key issue appears to be that the Mayo Clinic decided to stop using Prometheus's product and started doing its own testing; in response Prometheus argued that in using the separate Mayo test doctors would still "[think] about the correlations described in Prometheus's patent" and that this would itself constitute infringement. SCOTUS disagreed, asserting that the activities of measuring thiopurine metabolites and from that determining appropriate drug dosages was, in the words of Justice Breyer, "well-understood, routine, conventional activity previously engaged in by scientists who work in the field." And thus, unpatentable.

Unfortunately, although I believe this decision to be correct, I do not think the verdict is going to do much to untangle the present snarl of what is or is not patentable in the US. It has become increasingly clear in the past decade that decisions such as in re Bilski have done nothing to clarify what ought to happen. Instead, the waters have gotten more and more muddied.