From March to September 2016 the team is joined by Guest Kats Emma Perot and Mike Mireles.

From April to September 2016 the team is also joined by InternKats Eleanor Wilson and Nick Smallwood.

Thursday, 25 June 2015

Not so secret agent: when Bond isn't 007 but 0.77

In "The wounded patent survived, was only just infringed, but no injunction", here, fellow Kat Darren wrote about the decision of Birss J in Smith & Nephew Plc v ConvaTec Technologies Inc [2013] EWHC 3955 (Pat), a technically detailed case which amused Merpel, who commented that a case that started off being basically about chemistry ended up being basically about mathematics. Yesterday the Court of Appeal (Lords Justices Kitchin, Briggs and Christopher Clarke), at [2015] EWCA Civ 607 , allowed ConvaTec's appeal and dismissed Smith & Nephew's cross-appeal.

In short, ConvaTec's patent was a method of making light-stabilised silverised antimicrobial materials for the silverisation of gel-forming fibres in wound dressings. This was a three-step method, of which the third step involved subjecting gel-forming fibres to a binding agent. According to the relevant claim, this agent had to be present in a concentration of "between 1 per cent and 25 per cent". Smith & Nephew made its own range of wound dressings in accordance with method described in the patent, except that it used a concentration of binding agent of no more than 0.77 per cent -- which even a fictional cat can see is not in the range of 1 to 25. Said ConvaTec, this still infringed because, in a scientific context, the standard practice was to express numerical values to a particular degree of accuracy; using the standard rounding convention, the number five was always rounded up. On this basis "between 1 per cent and 25 per cent" encompassed all concentrations greater than or equal to 0.5 per cent and less than 25.5 per cent. No, said Birss J: the range was to be interpreted on the basis of significant figures, not rounding to the nearest whole number, so that the claim included concentrations greater than or equal to 0.95 per cent and less than 25.5 per cent (on that basis he held that, while the 0.77-based process did not infringe the patent, an earlier process did). This decision didn't seem to make anyone very happy, so both parties appealed.

On appeal ConvaTec argued that the range of "between 1 per cent and 25 per cent" embraced all concentrations greater than or equal to 0.5 per cent and less than 25.5 per cent, while Smith & Nephew said it encompassed all concentrations greater than or equal to precisely 1 per cent, and less than or equal to precisely 25 per cent.

Allowing Smith & Nephew's appeal and dismissing allowing ConvaTec's cross-appeal, the Court of Appeal (for whom Kitchin LJ delivered the judgment) held as follows:

* The approach to be adopted when interpreting claims that contained a range of numbers was the same as that to be adopted in relation to any other patent claim: the court had to be objective, asking what a skilled person would have understood the patentee to mean. The claims were to be construed purposively and in context [while the court cited more modern precedents, this Kat was pleased to see Catnic Components v Hill & Smith Ltd [1982] RPC 183 (HL), getting a mention: that's where "vertical" was generously construed as being a full six degrees off perpendicular].

* Where a numerical range was being construed, the question was what the person skilled in the art would understand the numerals to mean, in the light of common general knowledge as well as the context of the specification as a whole.

* Would that skilled person understand that the patentee had expressed the numerals to a limited degree of precision so as to include all values falling within the claimed range when stated with the same degree of precision? That would depend on all the circumstances, including the number of decimal places or significant figures to which the numerals were expressed.

Waiting for Bond ...
* In this case there were two questions to answer.  First, would the skilled person believe that 1 per cent and 25 per cent were exact values, or would he understand that the patentee had used a standard number convention to express the limits of the claim to a lesser degree of accuracy? Here, while the skilled person could measure concentrations of binding agent to a high degree of accuracy, it was clear from the teaching of the specification that the binding agent concentration was not critical -- and it appeared from the specification that the patentee was aware of the possibility of expressing numerical values with a high degree of precision. Accordingly the skilled person would understand that the patentee had chosen to express the numerical limits of the range to only a limited degree of accuracy. There was no great significance in the literal meaning of the word "between", which simply denoted the outer limits of the range, saying nothing about the degree of precision with which they were expressed.

* Secondly, would the skilled person understand the numbers to be expressed as whole numbers or in terms of significant figures? There was no logical basis for preferring significant figures approach [if you've ever wondered what is significant about significant figures, take a look at the quote in para 24: to the scientist there is a difference between 1.3 and 1.30 ...]. Here ConvaTec sought to convey the range of permissible binding agent concentrations and the accuracy with which those concentrations had to be determined. The skilled person would understand the 1 per cent and 25 per cent limits to have been expressed to the nearest whole number and the trial judge should therefore have found that the skilled reader would understand that ConvaTec intended the claim to embrace all concentrations of binding agent greater than or equal to 0.5 per cent and less than 25.5 per cent.

Book of Numbers here and here

14 comments:

Anonymous said...

NB The result of the appeal is described incorrectly in the 4.0th paragraph. In fact, Convatec's appeal (0.5-25.5%) was allowed and Smith & Nephew's appeal (1.0-25.0%) was dismissed.

Jeremy said...

Thanks, Anonymous -- a slip of the pen, as it were. Now duly amended!

Anonymous said...

This sort of thing annoys me. In Catnic (as I understand it), they had invented a frame with bits that went across and bits that went up, and claimed them as vertical and horizontal. They weren't considering cases in which the parts were nearly-but-not-quite vertical, and so there was a reasonable argument as to whether these should be included.

In the present case, the patentees knew that they were specifyng a number to indicate the end of a range - that the number was specifically there to say how far the claim went. One therefore ought to expect them to do so. Instead, we have a judgement saying that the number they actually meant is only half of the number they actually put. This makes a mockery of the requirement that the patentees should specify the protection that they seek, it makes a mockery of having an examination process, and it makes a mockery of the requirement that the claim construction should give a reasonable degree of certainty to third parties.

Brian Whitehead said...

Like Anonymous (22:43), we are having difficulties in understanding how the skilled person - a chemist - would regard 0.5% as falling within a claim specifying "between 1 per cent and 25 per cent". We think that the judges perhaps focused too closely on mathematical/numbering conventions, rather than approaching it from the point of view of a chemist, who would instead look at the issue in terms of margins of error. More at http://patlit.blogspot.co.uk/2015/06/precisely-delineated-claims-and.html

David Brophy said...

Posting on behalf of a commenter whose comment was rejected for unknown reasons by the Blogger platform:

May I suggest another way of looking at the matter?

The choice of the value 0.77 by S&N was anything but innocent. Its aim was manifestly to circumvent the claimed range.

What was disclosed in the patent was not only the range 1-2.5, but as well 0-1.

What S&N did was thus to make a selection within the range 0-1.

When applying the novelty criterion of selection inventions as found in the Guidelines under G-VI. 8(ii) one can decide upon the novelty of the selection. To decide positively, three criteria ought to be satisfied:

a) the selected sub-range is narrow compared to the known range; 0.77 is certainly narrow compared to the range.

(b) the selected sub-range is sufficiently far removed from any specific examples disclosed in the prior art and from the end-points of the known range; this point is debatable. For the sake of argument, let’s say that it is fulfilled.

(c) the selected range is not an arbitrary specimen of the prior art, i.e. not a mere embodiment of the prior art, but another invention. Here the notion of purposive selection (good morning Catnic), or of a new technical teaching through the selection play a role.

The last criterion is manifestly not fulfilled, and a hypothetical patent of S&N would lack novelty with respect of the available prior art. Hence infringement could be ascertained.

This way of looking into the matter is certainly daring, but is it not worth thinking of it?

Anonymous said...

This decision seems to be saying that “1 to 25” in the claim really means “0.5 to 25.5” and, following the same logic, “0.9 to 25” in a claim would really mean “0.85 to 25.5”.

So, if a claim started out as “0.9 to 25” and the patentee tried to narrow it during opposition to “1 to 25”, there could be an Article 123(3) attack, because the “true” scope of the claim has actually broadened (the lower limit moves from 0.85 to 0.5), even through the claim appears to have been narrowed from 0.9 to 1. This is clearly absurd.

The numbers in the claim were selected by the patentee to distinguish over the prior art and should never be rounded (or more precisely, as in this case, "reverse rounded").

What they should have done is say "0.77 written to one significant figure is 0.8, and 0.8 is less than 1, so no infringement".

Anonymous said...

The judgement concludes by stating that "the skilled reader would understand the patentee intended the claim to embrace all concentrations of binding agent >0.5% and <25.5%".

Does anyone seriously think that that ANY patentee/patent attorney having that intention would write a claim with a range other than "from 0.5% to 25.5%"?

Anonymous said...

Anonymous at 10:29,

Let's say the claim was from 0.5% to 25.5%.

Along comes a new party, and uses 0.49999999%.

Infringe?

Anonymous said...

Anonymous at 11:56:00:

In a sane system, no. If the patentee wants to catch people using 0.49999999%, he should write a claim with a lower limit less than or equal to 0.49999999%. Why should every number in a claim be changed, often unpredictably, to a different number which is "what they really meant"?

THE US anon said...

Same points of clarity I made earlier today on the Round thread apply with equal force here:


- the treatment of numbers in so far as "1" has a different meaning than "1.0" is NOT a new thing.
- that difference is very straightforward
- The decision "may" move goalposts - but in the opposite direction of which you speak and saves - not destroys - existing patents.

I do wonder if I am correcting the same person.

Anonymous said...

It would appear to me that people are taking the concept of the "claimed subject matter" and the "scope of protection" as being one and the same. They are not. a claim reciting between 1 and 25% is novel over a disclosure of 0.9% but adopting the reasoning of the Court of Appeal would most likely be non-inventive over this disclosure. A bit like the German "Formstein" approach - the claim cannot be interpreted to be read on to an arrangement which is not inventive over the prior art. Of course, absent any relevant prior art, the patentee is entitled to a broader scope of protection.

Darren Smyth said...

Thank you all for further and interesting comments. I feel moved to respond to a few. I am posting this on both threads relating to this Court of Appeal decision as there is some overlap with the comments.

To the US Anon:

I agree that “the treatment of numbers in so far as "1" has a different meaning than "1.0" is NOT a new thing”. However, as this case shows, it is not the case “that difference is very straightforward” – as explained in the post there is no settled caselaw (and in fact no case that I am aware of) that grapples with this question of whether the number should be interpreted according to number of significant figures or number of decimal places, which gives a different answer (in this case with infringement being found in the latter case but not the former).

Moreover, it is not the only consideration. There have been cases where courts have additionally or instead looked at the accuracy to which the measurement can be made, and those commenters who have referred to this aspect also raise an important and valid point.

I don’t see how you think that the decision saves rather than destroys existing patents – by expanding the range there runs the risk of it embracing prior art and therefore rendering them invalid. (To the Anon who says in this regard that "claimed subject matter" and the "scope of protection" are not one and the same – I do not believe that you are correct under UK jurisprudence. Once the claim has been construed, that construction applies to validity and infringement. So under the Court of Appeal construction, disclosure of 0.9% would be novelty destroying).

To Raoul in respect of your second comment:

You say that “The mere fact that it is possible to circumvent the patent by simply stating that the arbitrary value of 0.77 is not infringing makes me cringe.” The value of 0.77 is not arbitrary – it was selected to be outside the scope of the claim. You refer to “Art 69 and the Protocol on interpretation” – what about the part of that which refers to “reasonable degree of legal certainty for third parties”? Do you not think that third parties are entitled to work outside the terms of the claim so that they don’t infringe? I also don’t think it is fair to say that the courts came to diametrically opposite conclusions – they agreed on many things – that the claim did not mean precisely “1”, that rounding should be taken into consideration. They disagreed on the relatively minor issue of what rounding convention should be employed; it just so happens that this made the difference between infringement and non-infringement.

The prosecution history is interesting, but I don’t think it takes us much beyond what we conclude by looking at the claims – the 1-25% range has to be given a limiting effect. The question is only how much. Lord Justice Floyd has just reiterated that prosecution history should not be considered when construing claims (see http://ipkitten.blogspot.co.uk/2015/06/a-novel-becomes-saga-actavis-v-lilly.html).

Anonymous said...

Darren, while I agree that the UK has traditionally construed claims such that for both validity and infringement a common construction applies, this must be subject to question in the light of EP2000 and the A. 69 protocol - are you now allowed to argue that equivalents destroy novelty? I don't think so. They are, however, to be considered under Article 2 of the Protocol.

Darren Smyth said...

Thanks Anon at 11:27 - in the UK at least I think you are allowed to argue that the claim as construed including equivalents lacks novelty - I don't think that the EP2000 and its version of Art 69 Protocol has changed this. Other countries may take a different view. I am happy to be corrected if someone can point to contrary a decision. But the post EPC 2000 UK judgments all still seem to adopt the single construction approach.

Subscribe to the IPKat's posts by email here

Just pop your email address into the box and click 'Subscribe':