How do I know if a prior publication will invalidate my patent?
The Full Federal Court of 5 judges held in Bristol-Myers Squibb Company (BMS) v Apotex Pty Ltd  FCAFC 2 that a patent will be invalid for lack of novelty based on a disclosure in a previously published document (so called "prior art") if the prior art contains directions which, if followed, would inevitably result in the patented invention. But if more than one outcome is possible from the directions in the prior art, the patent is not invalid.
This case builds on previous case law which held that prior art will anticipate an patented invention if the prior art contains "clear and unmistakeable directions which, if followed, would result in the claimed invention." This is known as the "reverse infringement" test for patent infringement. The test is whether a previously published document has each and every one of the features of the claims of the patent.
In this case, the subject of the dispute was a drug called "crystals B" in BMS's patent. A prior art document contained instructions that resulted in the creation of Aripiprazole crystals. However, indifferent about the final step in the process of making crystals B, namely a drying step to obtain Aripiprazole in its free base form. A skilled person following the instructions in the prior art document would have had a range of options for this step.
While the method used by Apotex's expert resulted in crystals B, the Full Court held that it would have been open to a skilled person following the instructions in the prior art document to choose a different method, in which case the resulting crystals would not have had all of the attributes of crystals B. Thus, the result was not "inevitable" and the prior disclosure did not make the patent invalid.
The decision confirms that unless a prior art document contains a complete disclosure of the claimed invention, it is unlikely to invalidate a patent based on a lack of novelty.