My constant struggles with how the problem-solution approach is applied (or not)

Reading time: 10 minutes

Case No. O2018_004 | Decision of 14 December 2021

Part 1 of 2; to be cont’d in a separate post.

There is so much to write about this judgment that I hardly know where to begin. You might want to catch a glimpse of what is at stake on this Blog here. In my opinion, it is a judgment that would have deserved a headnote, maybe even more than one. A thorough discussion of all aspects will take some time. This is why I chose to take baby steps, to at least get things going. In this post, I will only address a tiny bit of the judgment concering the assessment of inventive step / problem-solution approach. A separat post will cover the other aspects of the decision, hopefully soon.

It hardly ever happens that I disagree with the assessment of inventive step in judgments of the FPC — as far as the result is concerned. But the reasoning sometimes puzzles me.

Case law should be predictable. Towards this end, it is comfy to have standardized tools at hand. This is where the EPO’s problem-solution approach comes into play (Guidelines, G-VII, 5) which the FPC has repeatedly indicated to apply. O2018_004 is no exception to the rule:

As a quick recap, the problem-solution approach involves three main steps:

    1. determining the ‘closest prior art‘;
    2. establishing the ‘objective technical problem‘ to be solved; and
    3. considering whether the claimed invention would have been obvious to the skilled person.

Even though case law should be predictable, individual judgments must be ‘right’. Accordingly, the judgment holds that the problem-solution approach shall not be formalistically but critically applied, to avoid hindsight bias, in particular in the formulation of the objective problem. I do concur (it’s even in the Guidelines, G-VII, 5.2, third but last paragraph).

The judgment goes on to explain why:

Losely translated to English:

If the effects or problems associated with the distinguishing features cannot be inferred from the closest prior art and are not obvious for the skilled person in light of the teaching of the closest prior art, the schematic application of the problem-solution approach can and usually will result in the identified ‘objective technical problem’ already giving an indication of problems or improvements in a certain direction which cannot be inferred from the closest prior art and which is remote therefrom.45

45 T 5/81 of 4 March 1982, OJ 1982, 249, ¶11; T 63/97 of 1 December 1997, ¶3.5; T 170/97 of 23 February 1998, ¶2.3; T 414/98 of 30 November 1999, ¶6.1.

That’s a looong sentence that I had to rehash several times to swallow it — and I am not sure if I’m done yet.

The decision goes on to hold:

In such situations, the objective technical problem must not be formulated in terms of the specific effects associated with the differentiating features; otherwise it would already give the skilled person a retrospective indication to even consider changes in a certain direction for improvements.

Well, … I disagree. It is perfectly fine in the problem-solution approach to include effects (Effekte) or problems (Problemstellungen) into the definition of the objective technical problem which are not disclosed or rendered obvious by the closest prior art. If that was not a given anymore, one would push the door gaping wide open for the test to deliver obviously wrong results.

Here is an example:

Let’s take a prior art document D1 that discloses a snow plow for mounting on a truck. The whole document is about the clever geometry of the plow itself. The mounting for connecting the plow to the truck is only shown schematically, not described in any detail. No discussion whatsoever about the mounting, or that there could be any issues with the mounting. Taken to the extreme: Let’s assume that D1 boldly holds that any mounting will be equally fine.

The later invention builds on the plow of D1, and improves the mounting. Let’s say, the inventive mounting for the plow is more strain resistant than any other mounting that had been known for snow plows of any kind. Let’s assume that this is due to rivetting instead of welding.

There is nothing wrong about a formulation of the objective technical problem to make the mounting of the plow according to D1 more strain resistant: If there is a secondary document or other teaching that fairly suggests a strain resistant mounting for a plow of the kind of D1, the invention is dead on arrival (rightly so). If not, an inventive step is to be acknowledged.

I would only see it differently in the rare case of the discovery of a previously unrecognised problem, so-called problem inventions. In the above example, let’s assume that previously known snow plow mountings rapidly broke and noone had ever understood why. If the issue of strain in comparable mountings had never ever been mentioned in the prior art, the formalistic appplication of the problem-solution approach would indeed lead to ‘wrong’ results. But, frankly, arguing for a problem invention typically is a desperate last resort, evidenced be a whole body of Case Law of the Boards of Appeal. What is more, posing of a new problem must be shown over the prior art as a whole (similar to a technical prejudice), not only vis-à-vis the closest prior art. And the burden is on the applicant / patentee to do so, not on the office or court to come up with it.

If it would happen once in a while that the problem-solution approach cannot / must not be applied as we all know it, then I would expect that this would be mentioned in the EPO Guidelines. But there is … nothing.

Next, the references to case-law of the Boards of Appeal in fn 45 (see above) made me worry whether I missed anything.

Let’s see:

T 5/81 of 4 March 1982 (OJ 1982, 249) is so old, it’s almost prehistoric. The decision even predates T 24/81 of 13 October 1982 which has outlined the problem-solution approach for the very first time. The reasoning in ¶11 of T 5/81 evolves around a comparison of the problem indicated in the closest prior art document and the problem defined in the application as filed. That had been one of many approaches at the time to tackle obviousness. But not so anymore today. T 24/81 in no way sets itself apart from the problem-solution approach.  

As to T 63/97 of 1 December 1997, I feel that the reference to ¶3.5 is besides the point. The definition of the problem (step ii., above) is already done in ¶3.4; ¶3.5 deals with the assessment of obviousness of the solution (step iii., above). I have a vague fealing that the reference is meant to point to ¶3.5.3, holding the following:

An interpretation of the documents as influenced by the problem solved by the invention while the problen was neither mentioned or even suggested must be avoided, such an approach being merely the result of an a posteriori analysis (see decision T 5/81, OJ EPO 1982, 249)

I do concur: Prior art documents must not be interpreted with hindsight. But this concerns the assessment of whether or not the solution had been obvious (step iii., above), not the definition of the objective technical problem (step ii., above).

Reference to T 170/97 of 23 February 1998 does not add much, at least at first glance. The objective technical problem is merely stated without any further ado in ¶2.3. What is interesting, though, is the reference to T 13/84, which is the decision that had introduced the concept of re-formulation of the problem in view of newly cited prior art. In T 13/84, the applicant argued that the preamble of the claim (reflecting the closest prior art) should relate to the same or a similar problem. The Board explicitly disagreed:

I have discussed a very similar issue on this Blog here in relation to the choice of the closest prior art, i.e. when it had been held in O2015_008 that the objective technical problem that is used for the problem-solution approach shall have a «basis or motivation in the closest prior art» document. I feel that T 13/84 supports my view that this is not correct.

T 414/98 of 30 November 1999 is besides the point, again. Reference to ¶6.1 concerns the assessment of obviousness of the solution in the problem-solution approach, i.e. step iii., above. The definition of the objective technical problem is being dealt with in ¶5, in no special way.

In sum, none of the references in fn 45 supports the view that there is a systematic issue with the problem-solution approach when «effects or problems associated with the distinguishing features cannot be inferred from the closest prior art.» If that was indeed an issue, it would almost certainly have been dealt with in quite a number of decision of the Boards of Appeal meanwhile: There are thousands of decisions that have applied the problem-solution approch, over a timeframe of four decades. But it appears that there is not a single precedent.

Now that you followed me through this lenghty exercise to establish a line of reasoning why this approach is mistaken, you might wonder:

Does it make a difference?

Not in the present case, I believe. The secondary reference is technologically so remote from the closest prior art that one could well have denied obviousness in step iii. of the standard problem-solution approach.

But that does not mean it is irrelevant. It significantly impairs predictability of case-law, without any apparent need.

What is more, it might well make a difference in another case, which it shouldn’t. The problem-solution approach is an extremely valuable tool to ensure predictability of case-law since about 40 years in thousands of Boards of Appeal decisions, and tens of thousands of decisions to grant or refuse patent applications at the EPO on (non-)obviousness — each year.

The problem-solution approach surely is not the one and only tool to assess obviousness. But since the FPC opted to apply it as a rule, I would hope that it is only modified with a trembling hand if there really is a need for it. Apart from problem inventions, I cannot see any.

UPDATE 14 May 2021:

The decision has been appealed.

Reported by Martin WILMING


Case No. O2018_004 | Decision of 14 December 2021

Synova AG
Avonisys AG
Stephan MOHREN

Panel of Judges:

    • Dr. Andri HESS
    • Dr. Tobias BREMI
    • Lara DORIGO


    • Dr. Tobias BREMI

Court Clerk:

    • Susanne ANDERHALDEN

Representative(s) of Plaintiff:

    • Dr. Michael RITSCHER (MLL)
    • Louisa GALBRAITH (MLL)
    • Werner ROSHARDT (Keller Schneider), assisting in patent matters
    • Stephan KESSLER (Keller Schneider), assisting in patent matters

Representative(s) of Defendant:

    • Dr. Ralph SCHLOSSER
    • Dr. Regula RÜEDI (Blum), assisting in patent matters
    • Dr. Christoph ENGELBRECHT (Blum), assisting in patent matters
    • Dr. Zacharias STELZER (Blum), assisting in patent matters



Enter your name and email address below to get notified of new posts by email.

Print Friendly, PDF & Email

One Reply to “My constant struggles with how the problem-solution approach is applied (or not)”

Leave a Reply

Your email address will not be published.