From its earliest days, the EPO has adopted a so-called "problem and solution" approach to questions of obviousness. As this has evolved over the years it has required an analysis to determine what is the closest prior art and then posed the question of whether an inventive step was required to go from that closest piece of art to the invention as claimed. Determination of whether that step is in fact inventive depends upon whether the state of the art would have led the notional skilled person in the art to the subject matter of the claims in suit. The characteristics of that notional skilled man were summarized in Alpha-interferon II/Biogen[1] where they were stated as being represented by a team of appropriate specialists, oriented towards practicalities but not solving technical problems by performing scientific research in areas not yet explored.[2] In this case, the board, after having distinguished earlier case law holding that the claiming of a purer form of a known substance to be a claim to a mere desideratum, upheld the patentability of a claim to "DNA sequences which hybridize to" any of a defined set of DNA sequences. The key to the applicants success was apparently the use of mixed oligonucleotides as probes to lead to obtain alpha interferon and in the early 80's there had been no expectation that such an approach might be successful.

There has been continuing criticism of the rigid application of the problem and solution approach over the years which recently resulted in an appeal board decision acknowledging that in certain circumstances other approaches might be more appropriate, particularly in cases where the prior art contained several different proposed solutions to the same problem as that solved by the applicant or where an invention breaks entirely new ground.[3] Nevertheless, the problem and solution approach is now sufficiently well established that it is likely to remain the standard test for inventivity in most cases.[4]

The issue of whether an application should be rejected or a patent held invalid on the ground that what is claimed was obvious to try has come to the fore recently in the biotechnology area. The EPO has, however, forsworn a simple obvious to try test in decisions such as AKZO/Dry jet-wet spinning,[5] in which the Appeal Board commented that, even in the absence of a prejudice in the art against taking the step claimed by the applicant, taking such a step could still be patentable if the art simply showed that:

In Expression in yeast/Gennentech[6] this principle was followed to revoke a European Patent covering use of recombinant techniques in yeast which had previously been used to express genes in bacteria where one skilled in the art would not have expected serious difficulty or risk in applying the teaching in one field to the other. Thus the application of the technology developed in bacteria to yeast was obvious to try with a reasonable expectation of success.[7] In Monoclonal antibody therapy/Ortho[8] it was affirmed that in addition to there being a reasonable expectation of success to reach a conclusion of obviousness, the reader of the prior document must also be able to have put into practice the suggestion in question without difficulty. It was, however pointed out that the mere fact that a document contains more than one suggestion for solving a problem does not prevent one from focussing on only one of these to conclude that what is claimed is obvious.

In Hepatitis B/Biogen, [9] however, where the Appeal board reversed an initial revocation of the patent, it was pointed out that one should not confuse "a reasonable expectation of success" with an "understandable hope to succeed". In this case the mere fact that other teams were striving to achieve the same object as the patentees did not establish any expectation of success in view of the unexplored nature of the field in which the inventors were working.[10] According to the Board, a reasonable expectation of success "implies the ability of a skilled person to reasonably predict on the basis of existing knowledge before the starting date of a research project, a successful conclusion to the said project within acceptable time limits".

A related case also pertaining to hepatitis B antigen has attracted attention in view of apparently conflicting decisions of the EPO Appeal Board (upholding the patent) and the House of Lords (holding it invalid). The House decided that the patent was invalid, essentially on the ground that it was too broad. It was accepted that there may have been an invention at the core of what the inventors had done. But what was claimed was in effect every way to solve an obvious problem. The claim defined the product claimed by its properties namely that it was recombinant DNA that set out the code for a protein having hepatitis B virus antigen specificity. In their lordships view it was not right to grant a patent that might stifle healthy competition by "allowing the first person who had found a way of achieving an obviously desirable goal to monopolize every other way of doing so". In such cases, the real invention was apparently the particular means devised to reach that goal and the claim should be framed accordingly.[11]

The question of whether by being cast too broadly may expose a claim to an obviousness attack that would not otherwise succeed has, however, also been recognized in the European Patent Office in, for example, Triazoles/Agrevo[12]. In this case, the examining division had held that a claim was so broad that it covered compounds for which it was not credible that they had the stated technical effect and that the specification was therefore insufficient in failing to describe how to obtain this effect. The Appeal Board found that this was not a valid ground for rejection as such but went on to hold that in a case where the question of inventivity was dependent on showing that compounds had a particular effect (in this case they were similar to prior art compounds so that the problem to be solved was selection of those compounds that had meritorious properties) then a claim that extended beyond those compounds that it was credible had the technical properties in question would cover obvious compounds and so be unpatentable because it lacked an inventive step.

In Benzodioxane derivatives/Esai[13] the question was whether it was obvious to replace chloro, bromo or trifluoromethyl substituents of 2-nitrobenzodioxanes of use in treatment of angina pectoris by cyano, carboxy or OR groups. The prior art reference was twenty years old. The applicants submitted no evidence of superior properties. In view of the lack of evidence of unexpected or superior properties, the Board concluded that the problem that the applicants had set out to solve was the production of alternative compounds for the same use as the prior art. This being the case, contrary to the applicant's submissions, the board would not discount the use of the prior document as the point from which obviousness had to be assessed on account of its age. The Board concluded:

The board was also unpersuaded that the fact that in some cases small changes in structure can result in large changes in biological properties would have dissuaded those skilled in the art from making the changes in question. The patentees had apparently not synthesized many of the compounds they were claiming and the board noted that in order to justify the patentability of such compounds an assumption had to be made that they had the same properties as those that had been made. There was therefore no reason to assume that the compounds claimed would not have the same properties as the prior art compounds and therefore be obvious over them.

As a practical matter, it should be noted that it is now difficult to persuade EPO examiners that a claim directed simply to recombinant DNA coding for a known protein is anything other than a claim to a desideratum.