[logosmonkey]

The biggest thing they do is file multiple patents for different parts of a drug. So they can have a 100 or so patents on a single drug and have the patents staggered over time. (Humira is a good example of this)

If that doesn't work they slightly alter stuff like dosage per pill and re-patent that.

[rayiner]

That doesn’t work because the original pill is itself prior art to every patent as of the date it’s released. So you either claim priority to the first batch of patents, or your later patent is defeated by your own prior art. Same thing with changing the dosage. Even if you can show a new discovery that wasn’t obvious from the original patent, that doesn’t prevent anyone from making a pill with the original dosage.

The way this works in practice is that there is a stream of patents on incremental improvements. Doctors continue to prescribe the branded, patented drug (because they have no reason not to) even though the original drug was fine.

Insulin is like this. If there was a substantial market for the 1923 bovine insulin produced by Eli Lilly, it would be cheap. Instead, insurance is willing to cover the high prices of newer, incrementally better versions. Leaving a relatively small market for a worse, but patent free version.

[wiggles_md]

You could, if you were dealing with a new chemical entity and were so inclined, file a new application with human pharmacokinetics claims, and then elect the PK patent for your FDA-assisted patent term extension. An excellent choice, since the PK claim patent will likely have the most recent priority date.

[xyzzyz]

Indeed, there is market for it, and it is actually very cheap ($25 for a vial in Walmart, which typically lasts around a month). The problem with it is that it’s much more inconvenient and difficult to use correctly. The modern analogues are much superior, so most people prefer to go for those instead, despite much greater price.