To what extent can a rule developed in the context of comparing numerical values of a patent claim and the numerical values of prior art be applicable to an internal comparison between examples and claims of a patent? The article below addresses this issue.
Common EPO practice involving numerical values when assessing patentability
Patent claims often contain ranges of values with upper and lower limits specified with certain precision. As an example, let us consider a claim with an upper limit of "1" (instead of, for example, "1.0"). Established case law of the Board of Appeal of the EPO teaches us that when assessing the novelty of a patent claim containing a numerical value (“1”) alongside prior art disclosing a numerical value with higher precision ("1.1"), the rules of rounding may be applied such that 1.1 may be considered the same as 1, resulting in the claim being considered not novel (see e.g. decisions T 1186/05, T 708/05 and T 871/08).
In contrast, according to the Boards of Appeal, the principle described above with respect to patentability should not be applied per se with respect to differences in precision between numerical terms in the claims and numerical terms included in examples in the European patent (see decision T 0942/17). In any event, it is good practice, when drafting a patent application and claims, to apply numerical values consistently and with precision.
Inventive step and experimental data
Patents often include examples consistent with the invention, as well as comparative examples that are not part of the invention. As an example, consider an invention relating to the content of component A in a range from 0.1 to 1%, which, according to the patent, produces beneficial effect B. The patent includes examples in which the content of component A is 0.1 to 1%, and comparative examples in which the content of component A is less than 0.1% or greater than 1%, that fail to achieve the desired effect B.
Examples: 0.1 % < A < 1% --> Effect B
Comparative examples: A < 0.1% or A > 1 % -/-> Effect B
Experimental comparisons of this type can show that certain characteristics (here the content of A) lead to a certain effect B. Unless it is known from the prior art that a particular content of A causes effect B, one may be able to formulate convincing arguments to demonstrate an inventive step. Suitable examples and comparative examples which are difficult to refute may therefore prove very helpful to the applicant/proprietor in grant or opposition (and appeal) proceedings.
Looking at the situation from the point of view of an attacking party in EPO opposition proceedings, a suitable strategy may be to demonstrate the opposite. For example, experiments that fulfill all features of the main claim but do not lead to effect B may prove particularly convincing.
Seemingly insignificant inconsistencies between examples and claims
Consider the following example in an opposition filed against a granted European patent.
A patent, with claims where content A is from 0.1 to 1%, contains a comparative example (comparative example = not according to the invention), which has 1.1% of content A. Since this is a comparative example, it does not show the beneficial effect B. At first glance, it seems obvious that 1.1% is more than 1%. But is this really the case? After all, content of A in the comparative example is indicated with higher precision (namely with one decimal place) than content A in the claim without a decimal place ("1 %" instead of "1.0 %").
If an opponent raises a corresponding objection, so that the Opposition Division and then the Board of Appeal address the following question: is the value of 1.1% within or outside the range of 0.1% to 1%? The opponent may look to related case law of the Boards of Appeal of the EPO dealing with the same question in a different context. This is because, as noted above, the case law long ago addressed the question of how numerical values from the prior art with higher precision (e.g., "1.1%") can be compared to numerical values from a patent claim having less precision: as discussed above, to enable meaningful comparison, the value from the prior art with higher precision must be adjusted to the precision of the value in the patent by applying common rounding rules, unless otherwise precluded.
If this rule from case law were to be applied analogously to our hypothetical case, the value of 1.1 % would have to be adjusted to "1 %" by rounding - with regard to the accuracy of the upper limit of the range "0.1 to 1 %". Thus, the value of "1.1 %" would be within the range "0.1 to 1 %". The example with 1.1 % rather than being considered as a comparative example as stated in the patent would be considered as an example according to the invention. Thus, because the 1.1% example does not achieve beneficial effect B, it would have to be taken as evidence that the beneficial effect B is not achieved over the entire range "0.1 to 1%". As mentioned above, the EPO rule in practice is often: "Without advantageous effect no inventive step, without inventive step no patent".
The result
Before we can reach a final decision in our example, the question still remains to what extent a rule developed in the context of a comparison between a patent and the prior art is applicable to an internal comparison between examples and claims of a patent.
At least in decision T 0942/17 (which is the basis for the simplified example discussed herein), the competent Board of Appeal answered this question in the negative. The Board found (subsequent values slightly adjusted) that by applying the rounding rules even a value that is almost 50% above the upper limit of 1% (for example 1.49%) would still be within the range "from 0.1% to 1%", so that the Board rejected the application of rounding rules to an internal comparison as inappropriate under the given circumstances. In addition, the designation of the 1.1% as a comparative example would confirm that an interpretation of the range "from 0.1 to 1%" using rounding rules was not reasonable.
For many, this decision of the Board of Appeal was welcome news. An unreflective application of rounding rules to an internal comparison between experimental data and claims of a patent can lead to circumstances where consistent experimental data of the patent proprietor are evaluated as evidence against the existence of an inventive step. Such Kafkaesque interpretation should be avoided whenever possible.
Even if this decision is in favor of the patentee, it is not necessarily impossible that in another situation a comparative example could be interpreted as an inventive example. For example, if the upper limit in the claim is not 1% but 9%, the application of the rounding rules does not result in a value that is almost 50% above the upper limit still falling within the claimed range. Here, it would only be about 6% (9.5/9 is roughly 1.06) so that a decision against the patent proprietor seems much more likely here.
Conclusion
When drafting patent applications, suitable examples and comparative examples should be formulated whenever possible. Such meaningful experimental comparisons can contribute to an uncomplicated grant procedure and strengthen the granted patent against attacks such as oppositions and nullity actions. Care should be taken to avoid inconsistencies between the scope of protection of the claims and the presentation of the invention in other parts of the patent application or patent, such as in the examples. In particular, one should avoid generalizing the subject matter of the patent claims to such a degree that comparative examples may be considered to be within the scope of the claims.
This article is also published in Lexology.
