Round, round: how to round? How do we round?

Please do not read this post if you have already read Jeremy’s beautifully succinct summary of the recent appeal hearing between Smith & Nephew and ConvaTec, and have absolutely no inclination to get further bogged down in numerical semantics. However, if you’re a glutton for punishment or have a (somewhat unhealthy) predilection for rounding conventions (as, strangely, this Kat does), then off we go.

ConvaTec Technologies and Smith & Nephew have added another court appearance to their long list of confrontations. On 24 June, the Court of Appeal handed down its decision in Smith & Nephew Plc v Convatec Technologies Inc & Ors [2015] EWCA Civ 607, an appeal of Birss J’s 2013 High Court judgment (reported by the IPKat amid New Year frivolities here), which can be read on BAILII.

The patent in suit was EP (UK) 1 343 510 in the name of ConvaTec, a patent for a light-stabilized silverisation of gel forming fibres in the context of an antimicrobial component of a wound dressing.

Mr Justice Birss’ Decision

The first hearing was instigated by Smith & Nephew as they sought a declaration of non-infringement for their product Durafiber Ag. Convatec counterclaimed that Smith & Nephew’s Durafiber product would infringe the patent. There were in fact two Smith & Nephew products to be considered: an earlier incarnation of the Smith & Nephew product ("the Original product") and an updated version (“the Modified product”); this Kat only concerns himself here with the events surrounding the Modified product, which is where the main issue of construction of numerical limitations in claims occurs.

Whilst the patent itself has an undeniably chemical flavour, the case turned on mathematical semantics. Claim 1 of the patent in suit (as amended in earlier proceedings) reads:

A method of preparing a light stabilized antimicrobial material, characterised in that the method comprises the steps of:
(a) preparing a solution comprising an
organic solvent and a source of silver in a quantity sufficient to provide a desired silver concentration in said material;
(b) subjecting a material which includes gel-forming fibres containing one or more hydrophilic, amphoteric or anionic polymers to said solution for a time sufficient to incorporate said desired silver concentration into said polymer, wherein said polymer comprises a polysaccharide or modified polysaccharide, a polyvinylpyrrolidone, a polyvinyl alcohol, a polyvinyl ether, a polyurethane, a polyacrylate, a polyacrylamide, collagen, or gelatin or mixtures thereof; and
(c) subjecting said polymer, during or after step (b) to one or more agents selected from the group consisting of ammonium salts, thiosulphates, chlorides and peroxides which facilitate the binding of said silver on said polymer, the agent being present in a concentration between 1% and 25% of the total volume of treatment, which material is substantially photostable upon drying, but which will dissociate to release said silver upon rehydration of said material.

The contentious issue of claim construction rested on the interpretation of step (c), and deciding what the skilled person would understand “between 1% and 25%” (as shown in bold above) to mean: are 1% and 25% the absolute limits of the range, or are they imprecise numbers which could be the result of ‘rounding’ conventions?

How to round up cats

ConvaTec argued that “between 1% and 25%” did not define precise values, but would be seen as whole numbers which could be the result of the basic whole number rounding convention; that is, 1% encompasses values ≥ 0.5% and < 1.5%, and 25% encompasses values ≥ 24.5% and < 25.5%. Therefore, the claimed range includes all concentrations of binding agent equal to or greater than 0.5% and less than 25.5%. According to this approach, Smith & Nephew’s product (which contains 0.77% binding agent) would fall within the scope of the claim.

Smith & Nephew, on the other hand, argued that the limits of the claimed range were precisely as they were stated (i.e. a concentration of 0.999% would not fall within the scope of the claim). This seems to go too far, and this Kat considers that such an argument was always doomed to failure.

However, their fall back argument was that the skilled person would understand the values to be correct to the number of significant figures displayed. That is, the number 1 is correct to one significant figure, and so encompasses all values ≥ 0.95 and < 1.5 (as a result of the “mathematical quirk” discussed below). The number 25 is correct to two significant figures, and so encompasses all values ≥ 24.5 and < 25.5. Therefore, the scope of the range claimed would be ≥ 0.95% and < 25.5%. This would mean that Smith & Nephew’s Durafiber Ag product, prepared using only 0.77% binding agent, would not infringe.

Birss J rejected the argument that the claimed range represented absolute limits as primarily argued by Smith & Nephew, but ConvaTec’s ‘whole number rounding’ argument did not convince him either; instead, he held that the numbers were accurate according to the number of significant figures displayed. Smith & Nephew’s Durafiber Ag product was therefore found not to infringe.

Court of Appeal Decision

Following the first hearing, both parties appealed against Birss J’s construction of the numerical range – although Smith & Nephew may appear to have achieved what they required in the first trial, their appeal was principally in relation to issues relating to the Original product.

ConvaTec once again argued that the claimed range represented whole numbers which could be the result of the basic rounding convention, and Smith & Nephew maintained that the limits of the claimed range were precisely as they were stated (this Kat is surprised that Smith & Nephew attempted this line of argument again). However, their fall-back position was that Mr Justice Birss was indeed correct - this Kat further wonders whether Smith & Nephew perhaps would have had more luck at the appeal if they had fully endorsed this aspect of Birss J’s decision.

The Court of Appeal considered a number of UK and EPO cases. Much of the UK case law indicated that the numerical values of claimed ranges should be interpreted to represent values correct to the number of significant figures displayed. On the face of it, this agrees with Judge Birss’ construction.

EPO case law featured a greater disparity of claim construction in similar cases, but the court highlighted EPO decision T1186/05 Multilayer films/Cryovac as a somewhat harmonising decision. Here, the EPO Board stated that, when comparing a patent which claimed a range of 0.89 - 0.92 to a piece of prior art which disclosed an amount of 0.885, the 0.885 value should be rounded to 0.89, as “an appropriate comparison [can] only be made with the prior art if its disclosure [is] rounded to the same degree of accuracy”.

Lord Justice Kitchen summarised some ground rules for numerical range construction in paragraph 38 of the decision:

“First, the scope of any such claim must be exactly the same whether one is considering infringement or validity.
Secondly, there can be no justification for using rounding or any other kind of approximation to change the disclosure of the prior art or to modify the alleged infringement.
Thirdly, the meaning and scope of a numerical range in a patent claim must be ascertained in light of the common general knowledge and in the context of the specification as a whole.
Fourthly, it may be the case that, in light of the common general knowledge and the teaching of the specification, the skilled person would understand that the patentee has chosen to express the numerals in the claim to a particular but limited degree of precision and so intends the claim to include all values which fall within the claimed range when stated with the same degree of precision.
Fifthly, whether that is so or not will depend upon all the circumstances including the number of decimal places or significant figures to which the numerals in the claim appear to have been expressed.”

Kitchin LJ firstly dismissed the argument that the limits should be taken as exact values. This decision was reached by considering that the patentee had in mind the possibility of expressing numerical values to a very high precision – the concentration of binding agent can be measured to a high degree of accuracy, and sections of the specification described concentrations accurate to two decimal places (i.e. 0.01%). It was also considered that the reader is taught that the invention can be performed with a binding agent concentration between 0.01 and 50%; the Court believed that this taught the skilled person that there is no technical reason to read the claim as requiring the use of a binding agent which falls between the exact numerical values of 1% and 25%.  [This Kat wonders what has happened to the doctrine that "what is not claimed is disclaimed" - there is surely an argument that if it is taught that the invention can be performed with a binding agent concentration between 0.01 and 50%, but what is claimed is narrower, then this narrower range should be given a strict interpretation.  Yet Kitchin LJ took the opposite view - "Once again this is a factor which undermines Smith & Nephew's favoured interpretation of the claim."]

Merpel is often prone to a little rounding

The Court then considered precisely how imprecise the patentee intended the limits to be understood.

As mentioned above, in comparable situations much of the UK caselaw directs a construction based on accuracy to significant figures, rather than a “whole numbers” approach. However, there is something of a “mathematical quirk” with this particular range claimed. As explained at the first trial, the range 1% to 25% using the significant figures approach encompasses numbers within the range ≥ 0.95% and < 25.5%. It is only the whole number 1 and powers of the number 10 to which such an asymmetrical range applies though; if the bottom of the claimed range had been 2% rather than 1%, even when applying the significant figures approach, the claim would have included all values ≥ 1.5 and < 25.5, a more broadly symmetrical distribution.

Counsel for ConvaTec explained further mathematical quirks with apparent aplomb (“helpful diagrams” were employed: Merpel can only hope that counsel managed to find a suitably feline number line). The patent description features a number of appropriate numerical ranges. One paragraph features the range from 1% to 10%, and another from 10% to 25%. Applying the significant figures approach to the first range would encompass all values ≥ 0.95 and < 10.5, but the second range would encompass ≥ 9.95 and < 25.5 (assuming that 10 is regarded to be accurate to two significant figures). As noted by Kitchen LJ, the error margin around 10% is an order of magnitude greater in one case than it is in the other.

The Court of Appeal felt that there was no reason to suppose that the accuracy conveyed by the patentee would depend on whether the bottom of the range is 1%, 2% or 5%, or whether 10% is at the top or bottom of the range. The Court of Appeal felt that Birss J had over-estimated the importance of the relative margins of error at the top and bottom of the claimed range, and under-estimated the importance of the anomalous figures which arise from the application of the significant figures approach (something of a judgmental rounding error).

Conclusion

The Court of Appeal therefore found that the claimed 1% - 25% range encompasses values of the range ≥ 0.5% and < 25.5% according to the “whole number” rounding convention. Hence Smith & Nephew’s Durafiber Ag product (and the method by which they create it) infringes ConvaTec’s patent.

This Kat was completely convinced by Birss J’s erudite reasoning and finds himself surprised that the Court of Appeal felt unable to endorse it.  He is troubled by a reasoning that considers that 1 as a lower limit of a range in a claim can actually mean 0.5.  From the comments on Jeremy's post, it seems that some readers are likewise uncomfortable with the Court of Appeal decision.  He was particularly taken with the point that both judgments (but the appeal decision more than the first instance) can be seen to focus on the mathematical conventions, rather than what a chemist would think about the claimed range. (This view is expanded in a piece on the PatLit blog, which also criticises the Court of Appeal decision in Actavis v Lilly.) What view do the rest of our readers take?