The Herchel Smith Intellectual Property Lecture 2022*

The UK courts' striving for consistency with international decisions in patent law
Richard Meade A judge of the High Court of England and Wales

Search for other papers by Richard Meade in
Current site
Google Scholar
Full access

The Patents Court, the Court of Appeal and the Supreme Court (before that, the House of Lords) have all and for many years stressed the need for consistency with decisions of the Boards of Appeal of the EPO. They have also taken account of decisions of the national courts of other EPC member states on many occasions in pursuit of consistency, in relation to validity but also infringement. This lecture will consider: how this has been done; when it has worked well and less well; how it can and should work if the EPO (in particular) changes or develops its law differently from the UK; and how that fits with the law of precedent. It will also consider the issues that may arise with decisions under the UPC.


The Patents Court, the Court of Appeal and the Supreme Court (before that, the House of Lords) have all and for many years stressed the need for consistency with decisions of the Boards of Appeal of the EPO. They have also taken account of decisions of the national courts of other EPC member states on many occasions in pursuit of consistency, in relation to validity but also infringement. This lecture will consider: how this has been done; when it has worked well and less well; how it can and should work if the EPO (in particular) changes or develops its law differently from the UK; and how that fits with the law of precedent. It will also consider the issues that may arise with decisions under the UPC.

I would like to start by saying thank you to Professor Gibson and Queen Mary University of London for asking me to give this Herchel Smith Lecture.

I am going to talk about how we try to achieve consistency with the EPO [European Patent Office] and with other European national jurisdictions when we decide patent cases. In a general sense this is well trodden ground, and indeed Colin Birss, now Lord Justice Birss but at the time HHJ Birss, touched on it in an earlier Herchel Smith Lecture, in 2012. But it has become more topical, and will present new challenges in the near future with the advent of the UPC [Unified Patent Court] and perhaps with upcoming decisions at the EPO.

Now to some considerable extent these topics are above my pay grade; I am just a judge sitting at first instance. I am bound by decisions of the Court of Appeal and of the Supreme Court. Any changes to the established UK approach would have to come from them, and later on in this lecture I will be looking at how the rules of precedent work if the EPO, in particular, gives decisions inconsistent with established UK law.

But working towards consistency does not always mean making big decisions or changing the status quo. It also lies in filling in smaller gaps, or making a decision on a new point. That of course is part of the work of a first instance judge. More generally, we all have an interest in looking for this kind of consistency, and understanding and commenting on how it is working, and how it should work.

I’m going to talk about the following things: the difference between working for consistency with the EPO and, on the other hand, working for consistency with other national jurisdictions; how working towards consistency means we can’t always have it our way; what we mean by consistency and how we assess whether we are achieving it; the effect of our law of judicial precedent in this area; some practical matters about how we do this work; and finally some looking into the future in the light of the UPC.

I’m going to take a few cases as focuses of my discussion: first, the decision of the Court of Appeal in Grimme;1 second, the decision of the Supreme Court in Actavis v Eli Lilly2 (the pemetrexed decision); third, the decision of the Supreme Court in Warner Lambert3 (the pregabalin litigation); and fourth, the decision of the Technical Board of Appeal of the EPO, the TBA, in T118/18, Sumitomo. That last is not perhaps such a household name in the UK patent profession as the others, but it is the case where the TBA referred to the Enlarged Board of Appeal some important questions about the role of plausibility in the law of patents, and in doing so commented on the Supreme Court’s judgment in Warner Lambert.

With some of those decisions and with others of the cases I am going to refer to, I must declare an interest: that I was involved as Counsel. I will also come clean and say that I wasn’t always on the winning side. Indeed, as I was pulling together the cases to put into this lecture it sometimes felt as if I never won any cases at the Bar. But I have tried to be objective, time dulls the pain of loss, and you are, if I may say so, a sophisticated audience able to filter out any sour grapes from what I may say.

Since I’m going to be referring to those four cases repeatedly, let me very briefly summarize them; in doing so I will be simplifying greatly.

In Grimme the issue was the test for contributory infringement. The defendant was selling potato cleaning machines, supplied with steel rollers and in which form they did not infringe. But some users might replace the rollers with rubber ones, and that then would fall within the claims. The issue was how to assess secondary infringement given the likelihood of replacement and the defendant’s state of mind about it. The actual test that the Court of Appeal reached is unimportant to what I have to say today; what matters is that the Court’s reasoning was heavily influenced by decisions of the German and Dutch courts on the same relevant provision of the CPC [Community Patent Convention]. Professor Sir Robin Jacob, then Jacob LJ, said:

the principle in our courts – and indeed that in the courts of other member states – should be to try to follow the reasoning of a court in another country. Only if the court of one state is convinced that the reasoning of another member state is erroneous should it depart from a point that has been authoritatively decided there.

In Actavis v Lilly, the alleged infringement used pemetrexed dipotassium when the claim language required pemetrexed disodium. Actavis sought a declaration of non-infringement under UK law and, by reference to the respective national designations of the EP in question, under the laws of France, Italy and Spain (and Germany, too, but that was no longer live by the time of trial). Arnold J, as he then was, and the Court of Appeal, said there was no infringement under any of those laws4 but the Supreme Court disagreed and said there was infringement under all of them, on the basis of equivalence. This involved a change in UK law over the previous House of Lords decision in Kirin-Amgen.5

An unusual feature of the case was that the Supreme Court was actually deciding infringement under the laws of other EPC [European Patent Convention] states as applied to the facts, not just looking to decisions of other EPC courts for guidance or for consistency.

In Warner-Lambert, the Supreme Court had to decide two major matters. The one relevant to my talk today was the law of sufficiency. After looking carefully at the EPO case law, it decided by a majority of three to two that a patent for a second medical use positively had to render plausible (to a modest standard) that the drug in question would achieve the treatment effect claimed.

In Sumitomo, the TBA was also considering validity of a medical use claim. In its decision it looked at the EPO’s own case law, and it looked in detail at the UK Supreme Court’s decision in Warner-Lambert. It identified strands of case law favouring what it called ab initio plausibility, and case law favouring what it called ab initio implausibility. It also touched on a third standard called ‘no plausibility’. At 13.5.5 it said that the difference between the first two of those ‘cannot be better illustrated’ than by the Supreme Court’s decision, in which the majority had preferred ab initio plausibility and the minority ab initio implausibility.

Recognizing that there was a divergence in its own case law and that the issues were fundamental ones, the TBA referred questions to the Enlarged Board. That reference is G2/21 and it will be heard in November.

The striking thing about the case is how it represents a dialogue, through their decisions, between the EPO and the UK Courts. The EPO rendered its decisions over some years about plausibility. The UK Courts looked at those and cited the EPO case law carefully and in detail in a number of decisions. The culmination came in Warner-Lambert. Now the EPO has shown that it was paying careful attention to what the UK was doing. And the dialogue will continue, because whatever the result, the Enlarged Board’s decision will be the subject of legal argument here quite quickly after it is given, no doubt.

So those cases are the backdrop for what I am going to say. Let me move on to the substance, and I will start by highlighting the central difference between looking for consistency with the EPO (focally in relation to validity) and looking for consistency with other EPC states’ decisions, which is much more to do with infringement.

The EPO first. A bedrock point here was made by Jacob LJ in Actavis v Merck6 at [47], beginning with a quote from the submissions of Counsel (Mr Prescott QC): ‘… you cannot have 34 [the current number of EPC members] ships steering in the same convoy unless there is something like a commodore’. Jacob LJ went on in his own words:

In saying our courts would and should normally follow the settled jurisprudence of the EPO it should be understood, of course, that they are not bound do so. In the unlikely event that we are convinced that the commodore is steering the convoy towards the rocks we can steer our ship away.

This followed a reference to Re Gale, from 1991,7 to the effect that a substantial divergence between the UK and the EPO would be disastrous because it would lead to a patent being granted or refused on the same facts depending on whether the application was made to the EPO or the UKIPO [UK Intellectual Property Office].
The Supreme Court has said much the same thing, in Eli Lilly v HGS8 at [87], in the judgment of Lord Neuberger (and more recently again in Actavis v ICOS9). Lord Neuberger said:

While consistency of approach is important, there has to be room for dialogue [that word again!] between a national court and the EPO (as well as between national courts themselves). Nonetheless, where the Board has adopted a consistent approach to an issue in a number of decisions, it would require very unusual facts to justify a national court not following that approach.

I would just elaborate by saying that a UK Court will only relatively infrequently have to give just a single decision of the EPO decisive weight unless it is a decision of the Enlarged Board. Decisions at the TBA level become important once there is a clear direction of travel, a consistent approach as Lord Neuberger put it.

So that is the EPO; we have a commodore, and an excellent one too. When it comes to the law of infringement in Europe, we do not have a commodore. There is no equivalent of the EPO. So Mr Prescott’s submission in Actavis v Merck that you have to have one wasn’t entirely right, although it was a powerful way of making his point. In relation to the law of infringement we don’t have a commodore but do have a convoy of dozens of ships. How do we make that work?

I’m reminded of a game my daughter played with her friends when she was little, about four or five. The game was called ‘leaders’. You can probably see where this is going. There were two classes of people in the game. One class was the leaders, and the other class was … everyone else. It was hard to find volunteers for the second class but my daughter had her own solution. She’d say ‘Let’s play leaders. You can all be the leaders and I’ll tell you what to do.’

In the no-commodore situation we cannot all be leaders all the time. Sometimes you have to be willing to follow. In Grimme, the Court of Appeal thought the German and Dutch decisions were right, but the flavour is very much that the Court would have followed them even if there were other possible ways to go.

In any case, while lacking a commodore we do have the CPC in common, and we have the Protocol to Article 69 of the EPC, which tells us that the ships in the convoy must set a course between two bounds: don’t use the claims just as a guide, but don’t be too literal, and have regard to equivalence. So the convoy has to head somewhere between, say, West and North.

In this no-commodore-convoy there are two things we can do. First, the ships can signal each other what course they think they should be plotting in future. That can be achieved, for example, by discussion among practitioners and judges. More of that when I discuss practical aspects.

Second, the ships can watch what the others have already done to interpret what course they think they are on and have been plotting and why, and on the assumption that they will, all things being equal, stay on it. This can be achieved by looking at actual judgments already given in other jurisdictions. This happens constantly, of course. There are feedback loops in both these approaches and they are not entirely distinct from each other.

I have already said that achieving consistency will sometimes mean following a path which would not necessarily have been the UK court’s first or intuitive choice, but which is a reasonable one and which is already adopted elsewhere. Let me give a couple of examples.

The first is from Kirin-Amgen, though not in the parts reconsidered by the Supreme Court in Actavis v Lilly.

The issue was that Amgen had a product by process claim to erythropoietin made by recombinant means: rEPO. The defendent argued at trial that the claim lacked novelty because rEPO, though made differently, was not physically different from erythropoietin purified from urine: uEPO.

In the EPO, the patent had survived this attack; the Board had held that the different ways of creating the erythropoietin necessarily did make them physically different. But the trial judge in the UK, Neuberger J, as he then was, came to a different conclusion on the facts. He held that the claimed products were not necessarily physically different by reason of how they were created. So the question was: for a product by process claim, is novelty conferred by the method alone?

Neuberger J’s conclusion was that it was. One can understand this intuitively: the claim was to erythropoietin made by a recombinant method, uEPO is not made by a recombinant method and so it cannot fall within the claim and cannot anticipate. The Court of Appeal agreed (led by Aldous LJ, one of the most experienced patent judges ever). This maintained the approach to product by process claims that the UK had had since even before the EPC.

The EPO position on this kind of claim was and still is somewhat subtle: product by process claims are allowed, but only when it is not possible to define the physical characteristics of the product other than by the way it is made. Neuberger J and the Court of Appeal said that this was a rule of practice and procedure and not a rule of substantive patentability.

The House of Lords disagreed: Lord Hoffmann said that in addition to that rule of procedure and practice, the EPO had a substantive rule that the product in a product by process claim had to be physically new. And although the EPO had allowed the claim in question, that had been because, on their findings of fact, rEPO was different from uEPO. Once it was found that they were not different, the claim was anticipated under the EPO approach.

I would suggest that that approach is also a very reasonable one: such a claim is a product claim, the ability to re-patent existing products should be carefully defended against, and the patentee could always have a method claim, so its interests were protected in relation to what it had invented.

For the purposes of what I am exploring today, the point is not about which intuitive view from first principles is better. The point is that the EPO had firmly chosen one over the other and the trial judge and the Court of Appeal had gone in a different direction. That had to be corrected. The EPO view was perhaps, even probably, different from the most natural view to a UK patent lawyer and the UK approach prior to the EPC. But if you are going to go for consistency you cannot always have it your way.

Another illustration may perhaps be found in the Generics v Lundbeck escitalopram litigation. The patentee claimed the more active of two enantiomers of the drug concerned; it had been appreciated in the art that one would be better than the other, but without a method of separation of the racemic mixture of them it was not possible to tell which. The patentee came up with such a method, but claimed not just the method but the better enantiomer as a product per se.

Kitchin J, as he then was, held that the claim was insufficient.10 Referring back to the decision of the House of Lords in Biogen11 he said that the technical contribution was only the method of separation not the product, and that a claim to the product made by any means exceeded that contribution.

Lord Hoffmann, sitting in the Court of Appeal, reached the opposite conclusion.12 He disagreed with Kitchin J’s analysis of Biogen, although it is important to note that he recognized the intuitive appeal of the first instance decision. But Lord Hoffmann also referred to a line of EPO cases (starting with T595/90 Kawasaki Steel Corporation13) which said that even for a known and obviously desirable product it was permissible to have a claim to it as such, if it could not previously be made and the patentee had come up with the first method to make it. As far as I can tell, those EPO cases were not relied on at trial.

Leaving aside the question of what Biogen decided as a matter of precedent, it seems that of two possible ways of thinking about the situation, each of which had intuitive appeal and arguments in their favour, only one would be consistent with the EPO’s case law.

So if we want consistency we cannot always have our own way. Let me move on to talk about what we mean by consistency.

The legislative sources of patent law that I am discussing are framed at a very high level of generality and the texts themselves don’t provide answers to the questions that we struggle with. What really matters is the case law. Take as an example, Article 83 of the EPC. It says the following ‘The European patent application shall disclose the invention in a manner sufficiently clear and complete for it to be carried out by a person skilled in the art’. It is sometimes objected this does not mention plausibility. But it also doesn’t mention sufficiency across the scope of the claim, or identify what we call Biogen insufficiency versus what we call classical insufficiency. I could go on. The real meat of our law comes from judicial decisions made against that very bare bones provision, and with reference to some fundamental notions of patent law such as the patent bargain.

Similarly, Article 56 dealing with inventive step just says that an invention shall be considered to involve an inventive step if not obvious. How to assess this fundamental legal requirement has been developed in many directions by the case law: obvious to try, Agrevo obviousness, commercial versus technical considerations, secondary indicia, and so on.

I do not suggest this is a problem and no doubt it was unavoidable given the way the EPC was negotiated, but it means that case law is where the substantive law of patents is to be found.

As a practical matter, it has to be recognized that perfect consistency will never be achieved, especially in a no-commodore model. If we had had from the start a system with very detailed legislative provisions so that every jurisdiction had to apply the same rules that would be one thing. But we didn’t and we don’t and differences were bound to emerge, and have done. We are in a situation on infringement where different jurisdictions have different approaches even on so fundamental matters as claim scope. You can see this from the Supreme Court’s decision in Actavis v Lilly. Whatever you think of the decision they made in relation to the UK, one can see from their summary of the position in other jurisdictions that France, for example, was appreciably different from Spain which was different again from Italy. Not outside the boundaries set by the Protocol, but within them, and some closer to one edge or the other. In this situation a change by one jurisdiction won’t achieve consistency; it can’t. But I would suggest that small changes towards what might loosely and probably wrongly be called the mean, the middle path, could help.

Another reason why perfect consistency is unachievable is that the challenges in patent law across Europe are not static. They change all the time. At the moment and for some years now FRAND has been a dominant issue; at the start of my career it barely existed. Patenting strategies change. New technologies come along – AI for example. Societal norms and attitudes change. All of these things imply developments for patent law and as courts of multiple jurisdictions start to deal with fresh challenges as they arise there will be some consistency and some inconsistency. It is a process and not a destination. It means there will always be a special responsibility for the first courts to look at a new area if others are going to take their lead from them.

So we have the commodore and no-commodore situations to consider; the case law is key; and we recognize that perfect consistency is not possible. But what actually is consistency in this context? When we say we are trying to achieve consistency with the EPO or with the courts of other EPC states, what do we mean?

Well, first of all we certainly don’t mean that we are trying to achieve a situation where the same case always leads to the same result when it comes before more than one European patent tribunal, be it national courts or the EPO, even if they are applying the same test. It has long been recognized that that is impossible, for at least three reasons. One is that the evidence and arguments may be different in different tribunals. A second, related to the first, is that the procedure used may differ (for example whether there is cross-examination, whether there is disclosure). A third is that when different tribunals do apply the same legal test, they can have different general inclinations in doing so. For example, some tribunals are felt to be stricter on added matter than others, and likewise on obviousness. This kind of divergence may be reduced over time by dialogue between tribunals or by reading one another’s judgments, or by listening to feedback from litigants and their advisers. But it cannot be captured by describing the test being applied, because the tribunals are applying the same test as described in their judgments.

So we can reject always getting the same result as being the goal. It is not possible. That doesn’t mean that when different tribunals reach different results in the same case we cannot learn from it. We may see from reading their judgments that the difference is attributable to applying different tests and not from mere differences in the evidence they heard. And we can also test consistency by asking how different tribunals would decide on a given set of postulated facts: thought experiments of the kind we are all familiar with.

What I would suggest we do mean by consistency between tribunals is their reaching approaches which are capable in most if not all circumstances of giving the same result, and although this goes without saying, which are consistent with the EPC or the CPC as the case may be.

Given that we are simply not going to adopt identically phrased or structured tests for implementing the broadly-phrased requirements of the EPC (and indeed the CPC), how do we go about, as we must, working out whether different tests are consistent? This is a key point.

I think the first thing to say here is: beware of the superficial. It is obviously tempting if you are assessing whether and to what extent, for example, the Supreme Court’s decision in Actavis v Lilly is consistent with the German position on equivalents, to jump straight to the three questions laid down for use in each jurisdiction. If you do that you will see what looks like a lot of consistency: a separate doctrine of equivalents on top of literal or normal interpretation; the use of three structured questions; textual similarity of the questions at each stage. Even at that level, I will be suggesting a little later, there are in fact real differences, but if you just compare the words of the questions themselves you will not be thinking about deeper issues like: what is the role of the claims?

So operating too superficially in making a comparison could lead you to think those tests are more similar than they really are. It could also lead you to think that tests which are in fact highly consistent, aren’t. Take the Pozzoli approach, which we use, and the problem solution approach (‘PSA’) of the EPO. They are phrased very differently, but repeated decisions of the courts here have established that in substance they are not. They just come at the same, single underlying statutory task in different ways. Aside from the analysis of the way the tests work, it is worth saying that there is a really striking lack, I think perhaps a complete lack, of cases where Pozzoli and the PSA would compel, or have compelled, a different conclusion on the same facts. Different conclusions have often been reached in opposition proceedings and in later national litigation, but this has always been because of the kind of factors that I referred to earlier: different evidence or procedure.

So it is not good enough just to compare the summary statements of the tests between for example the three questions of the Supreme Court in Actavis v Lilly and the German courts’ questions. But comparing them can legitimately be part of the picture.

For example, the Supreme Court modified the UK question 2 so that it now falls to be asked with knowledge that the variant does indeed work. German question 2 is addressed to what is sometimes called ‘discoverability’. It is somewhat subtle, but it does not build in automatically the assumption that the skilled reader knows that the variant works.

Are these consistent? They don’t look all that different at first. Question 2 only arises if question 1 has been answered in the patentee’s favour so that the variant is in fact not material, and question 2, at the most general level and paraphrasing very liberally, is about whether one is dealing with an obviously immaterial variant.

A real world assessment to test whether the respective question 2s are consistent is, I suggest, possible. One can simply ask: do the respective question 2s determine the result of cases in real life? And if so, how?

Well, in the Patents Court there have now been a considerable number of cases where the new question 2 has been looked at. I cannot think of any where the alleged infringer has won. Basically that is because as soon as the skilled person knows that something does work they also know or can immediately work out how. All relevant doubt is removed. So it seems that only in extraordinary circumstances will question 2 enter the picture separately from question 1 in the UK.

In Germany the picture is different. I was curious, when I came to write this lecture, about the real world role of question 2 in their courts. I cannot claim it is a scientific study, but I asked a former colleague from the English IP community to ask his partners in Germany for examples of cases where question 2 had gone against the patentee. They14 needed little time in coming up with ten or so examples where it had, from the Courts of Dusseldorf, Karlsruhe and Munich. They are just examples. There may have been other additional reasons why the patentee would lose anyway and there are other confounding factors, but this confirms, I think, that we find ourselves in significantly different places when it comes to equivalents.

So on question 2 there seems to be a fairly plain difference between ourselves and Germany despite apparent similarities in the respective questions. There probably already was some difference at the question 2 level when we used the Improver15 questions here, but I would suggest not as sharp and not as significant. It should be noted that the Supreme Court in Actavis v Lilly thought there would be consistency with Germany when it came to question 2 (at [62]). That may speak to the difficulty of predicting the effect of a decision when looking at just a single set of facts, but I think one can confidently say that on the materials we have so far, five years on, the prediction was amiss.

The comparison on questions 1 and 3 between ourselves and Germany is more analogue, more shades-of-grey, but certainly the Supreme Court’s decision in Actavis v Lilly has moved the dial, because although not reflected explicitly in the summary words of the questions, they were expressed against the background of a very significant reduction of the role of the claims compared with the UK position under Kirin-Amgen (see eg [60] and [65]). Most notably, under question 3 it is no longer permissible to take the words of the claims themselves on their own as excluding variants, even if the terminology in them is of a kind normally used for high or absolute precision (such as a chemical name). This has led to very few successes for the alleged infringer on question 3 in UK cases: really only where the kind of disclosed-but-not-claimed argument, known to German law, can be made. Before the Supreme Court in Actavis v Lilly, and in Germany before and since, there were and have been quite a fair few cases where the patentee lost simply because he or she was taken to have meant what was said by the claims.

Having said earlier that superficiality is to be avoided, I would also say that analysing consistency requires thinking not only about the point specifically in issue but the patent system more broadly.

What I mean is that the overall balance between patentee and defendant is not just dependent on infringement but on validity too. If you expand the scope of protection beyond the claims on their purposive interpretation, what do you do about the principle that you cannot re-patent the prior art? The Supreme Court’s decision in Actavis v Lilly raised the concern that patentees would find themselves with a broader claim scope than they had wanted or intended, and therefore in unforeseen and possibly fatal collision with the prior art.

One way around that could be to say that equivalents do not enter the picture for novelty – a party attacking the patent has to aim for the ‘normal’ claim scope. Perhaps that is where we will end up and there is obiter support for it in some decisions since, but it raises deep questions about what is sometimes called the golden thread: that you cannot re-patent the prior art.

The German system has a means for addressing the issue – the Formstein defence. But it isn’t unproblematic and it’s relatively complicated. Its complexity in part comes, as I understand it, from the bifurcated system. In German infringement proceedings it can be raised where the prior art that would fall within the scope of equivalents is in some way specific to the alleged infringement; other more general attacks that the patent is invalid on a broader interpretation have to be resolved in the federal patent court. That may move the golden thread issue to a different tribunal without entirely solving it.

There is also some obiter support for the introduction of a Formstein defence in the UK in the decision of Birss LJ (sitting at first instance) in Facebook v Voxer [2021] EWHC 1377 (Pat). It’s comforting that matters are developing towards a possible solution, but until that falls into place it is perhaps not even possible to say whether and to what extent, looking at patent law as a whole, the UK and Germany are more, or less, consistent as a result of Actavis v Lilly. If we stick with our new doctrine of equivalents without resolving the golden thread issue in the same way as Germany then we won’t be consistent in patent law overall.

Talk of Actavis v Lilly provides a segue to my next point, which is what happens when one of the ships in the convoy changes course? Clearly the Supreme Court in Actavis v Lilly was not saying that we should abandon the claims to the extent of using them only as a guideline. It was expressly addressing itself right from the outset to the balance required by the Protocol to Article 69 in its amended form (see [31] to [34]). But at the same time, because of the curtailment of question 2 and the other matters to which I have referred, the UK has become one of the most, if not the most, patentee-friendly jurisdictions in Europe when it comes to claim scope for infringement by equivalents. It is the role of the Supreme Court to make changes to the law of that scale; there are many who think that the modified test that the Court created is the right one, and it is certainly not my role to say whether it was right or wrong (I have been applying it in my judgments!). Nor is its desirability the subject of this lecture.

On the other hand, it is the subject of this lecture to use Actavis v Lilly as an example to think about consistency between the UK and other European jurisdictions. What is the effect on consistency of a major change of direction by one of the major jurisdictions? Well, it is obviously a challenge. Earlier, I said that in a convoy with no commodore the ships can talk about where to go, and each can look at the course other ships have been following.

In the context of using structured questions to address equivalents, other major European jurisdictions have followed our course over many years since Improver. Germany has questions similar to but not quite the same as Improver; Spain has exactly or almost exactly the Improver questions, as the Supreme Court in Actavis v Lilly noted (at [49] and [50]). To a significant extent they have followed this country’s earlier lead, therefore. It cannot easily be predicted, I would suggest, how readily they may take to a decision that the Improver questions were not the right answer and indeed were positively wrong.

When deciding the concrete question of whether there was infringement under Spanish law in the case before it, the Supreme Court in Actavis v Lilly accepted that Spain had previously applied the Improver questions, and (at least implicitly) that based on them there could be no infringement under Spanish law (as Arnold J had found). But the Supreme Court held (at [97]) that the Spanish courts would, if called on to decide infringement in relation to pemetrexed, change their approach and modify Improver so as to follow the UK’s new course. So it held that Actavis infringed under Spanish law, or perhaps one might say: Spanish law as the Supreme Court thought it ought to be.

Part of the reason for this conclusion by the Supreme Court was its analysis that it was moving the UK closer to the German position, a prediction which, with hindsight, cannot be said with confidence to have come true, as I have mentioned above. The higher Spanish courts have not, as I understand it, had to decide whether to follow our Supreme Court in Actavis v Lilly. As I understand it, the attitude of the lower courts is likely to be that they are bound by their own Supreme Court and only it can make the decision. One can understand that. Until there is such a decision, therefore, we do not know whether we are consistent with Spain or not, and that’s the result of our change of course.

So there clearly is a cost in terms of consistency associated with a major change of course of this kind; I would suggest that the bigger the change, the longer the existing course has been followed here, and the more it has already been followed by other jurisdictions, the higher that cost will be.

There is another aspect to this, perhaps the other side of the coin, which is expecting too much of a harmonizing or consistency-seeking decision. Changes in the law from the higher courts inevitably come with some wider impact, or are made without deciding, because it would be an impossible task, what all the consequential effects may be. I call this whataboutitis. So for example it’s sometimes said that Actavis v Lilly dealt with chemical names but what about numerical ranges? You can always pick holes like this. A degree of development by first instance judges in the light of a decision by the Supreme Court or the Court of Appeal is quite normal, healthy and necessary. Some appellate judgments expressly say that it is for the lower courts to refine their approach. For example in Virgin,16 the case about res judicata and damages when a patent is revoked by the EPO after being held infringed in the UK, the Supreme Court recognized that the approach to UK stays pending EPO oppositions would require review, and it followed in short order in IPCom.17 Similarly, when the Court of Appeal confirmed the discretionary nature of the Arrow declaration in Abbvie they said that it was for the Patents Court to develop the factors which would inform the exercise of the discretion.18

So big decisions will leave areas for future development where the position isn’t clear and where you therefore cannot say there is consistency, or not. It is plainly up to the higher courts to decide how much to cover and what to leave, and I am sure it is not an easy task. It isn’t clear why the Supreme Court left open the interaction between their decision on infringement and the law of claim scope for validity over the prior art. They were not charged with deciding validity between the parties because it was only a case about infringement; validity of the foreign patents in issue could not have been tried here, of course. On the other hand the arguments before the Court over claim scope for infringement were very much about the fact that broad claims are often bad for patentees, and neither intended nor desired by them, and it was a significant part of the reasoning in Kirin-Amgen. It has been a slightly disorienting few years not knowing if the claims have the same scope for infringement and validity, but even with such a major change we are working towards a settled position, as I have already discussed.

So I’ve said a fair bit now about consistency, that it is desirable for us to shape our law to follow the commodore, and that if you want to achieve consistency you cannot have everything your own way. These things mean that sometimes the UK might want to make a change in its case law. There, we have to think about the rules of precedent that the UK has.

In the simplest terms, the Court or Appeal is bound by its own decisions and by decisions of the Supreme Court. The Supreme Court is bound by its own decisions unless it chooses to depart from them according to its own rules originally set in a Practice Direction issued by the House of Lords. The High Court is bound by both. There are exceptions to the Court of Appeal being bound by its own decisions, a key case being Young v Bristol Aeroplane [1944] KB 718.

That is all fairly easy to follow, but what is to happen if our domestic law has the goal of following the decisions of an international tribunal, in particular and for present purposes the EPO, and the EPO produces case law inconsistent with binding domestic decisions?

This was considered by the Court of Appeal in Actavis v Merck [2008] EWCA Civ 444. The detailed facts do not matter; suffice it to say that it was a patent for a dosage regimen in a second medical use patent for a known drug and the party attacking the patent said that the dosage limits could not confer novelty as a matter of law. More than that, they said that there was a binding decision to that effect in a Court of Appeal case, BMS v Baker Norton [2001] RPC 1.

The patentee’s response included the argument that the dosage limits in fact were regarded as capable of providing novelty under EPO case law, in particular a TBA decision from 2006, that is, from after the UK decision in BMS.

The Court of Appeal held that in fact BMS didn’t contain a ratio that novelty could not arise from the dosage limits. But it also carefully considered what would have been the position otherwise: if there was a Court of Appeal authority on point and a later EPO decision, or line of cases the other way.

The Court of Appeal concluded that in those circumstances there needed to be, and it could create, a fresh exception added to Young v Bristol Aeroplane, arising from the EPC. It is that the Court of Appeal:

is free but not bound to depart from the ratio decidendi of its earlier decision if it is satisfied that the EPO Boards of Appeal have formed a settled view of European Patent law which is inconsistent with that earlier decision. Generally [the Court of Appeal] will follow that settled view.

And the Court contrasted the ECHR where the foreign tribunals and their decision-making methods are different.

The existence of this new exception was later confirmed by the House of Lords.19 It was a very different context – a judicial review about disability benefits where the main issue was consistency with decisions of the European Court of Human Rights – but one of the cases the House considered was Actavis v Merck. Lord Walker said20 the exception to Young was rightly created, although he also said the Court of Appeal had not been right to say that it could depart from a Supreme Court or House of Lords decision under the new exception. Speaking for myself, I don’t think the Court of Appeal did say that, but anyway the upshot is that if the UK law overtaken by EPO case law is a Supreme Court decision, only the Supreme Court can reverse it.

This is not a merely hypothetical scenario. As I mentioned earlier, in Sumitomo the TBA has referred to the Enlarged Board what is essentially the issue on plausibility that divided the Supreme Court 3:2 in Warner Lambert. It is clearly possible that the Enlarged Board might agree with the minority in the Supreme Court and relax the requirements of plausibility. If that happens then given the rules I have just been discussing it will be open to the Supreme Court, but only the Supreme Court, to change the UK position in response. Pending a suitable case for them to do that, lower courts will be bound by the decision in Warner-Lambert, although I do not see that rules out our proceeding on the basis in such a situation, for pragmatic purposes, that while we are bound by it, there is a non-theoretical chance of the law changing in the near future.

That is the situation with the EPC. The situation with infringement cannot be said to be the same. The new Young v Bristol Aeroplane exception does not apply to decisions of other national courts on infringement and anyway if they are not all consistent the justification for departing from settled UK case law is absent. So in that context the rules of precedent remain. Of course, the Supreme Court remains able to change its case law and in substance at least it did so in Actavis v Lilly by its departure from Kirin-Amgen.

What I have said so far is all quite theoretical. But looking for consistency is an intensely practical exercise. It has to start with finding what the law of the EPO and/or other relevant foreign tribunals in fact is. That exercise differs very markedly between the EPO commodore situation and the infringement no-commodore one.

Starting with the EPO, one has a single body of case law. Nearly all important decisions are available in English. The decisions are not usually especially long in their legal parts; the Boards tend to spend a lot more time on the facts. They are accessible online. I get TBA decisions in my regular Westlaw alerter and one can also find them easily on the EPO website.

There are also two excellent texts to use: the Guidelines for Examination and the Case Law of the Boards of Appeal (‘White’ Book). Not only are they very useful, but they are key, especially the white book, in working out what the overall direction of travel is and whether a proposition of law has become generally accepted at the EPO or is contentious.

On top of that, many UK patent practitioners have appeared at the EPO in oppositions and/or worked on cases where there are parallel EPO proceedings, and over the years, UK judges have also served on the Enlarged Board of Appeal, most recently Arnold LJ last year in G1/21, the ViCo dispute.

Taking all this together, one has reason for confidence – I feel confident that if I need to know the full and accurate picture of what EPO law is, then with the assistance of the parties and their Counsel, I can. Of course, there are some areas where EPO law is not settled. That’s a different point. If it’s not settled, I am in a full position to know that.

The position with decisions of other national courts in relation to the law of infringement is not as simple. This is not anyone’s fault; it is just life. One is dealing with dozens of different laws. Translation of the subtle points at play is not easy. The systems can be unfamiliar, including their systems of precedent (a point borne in on me very much during Actavis v Lilly with the Italian system – not a criticism of how they do things, just a comment on the task that faces one as a recipient of their decisions).

So it is rather harder to be fully confident that you have the absolute, complete and accurate picture.

This is not assisted by the fact that we do not have a developed procedural system for regulating the introduction of foreign decisions in our proceedings when they are being used simply as informative comparators. If one is seeking to prove what foreign law actually is then there have to be pleadings and experts reports, which entails notice to the other side and so on. Foreign decisions merely for persuasive effect are often just provided at trial in the authorities’ bundles.

I don’t at all mean to paint a bleak picture of this, though. There are three main positive things I would say.

First, this all arises in a situation where we know we aren’t going to get perfect consistency. If we have key decisions from two or three major jurisdictions and we work from them then we will achieve consistency with them at least. That was the situation in Grimme. The Court of Appeal had decisions from Germany and the Netherlands. It was able to make the first UK decision on the point at issue knowing that it would be in line with them at least.

Second, there has always been and will continue to be active dialogue in print and in person between the national judges. For example, in preparing this talk I have been greatly assisted by the many articles in the Festschrift for Peter Meier Beck published by GRUR;21 last week I participated in an EPO meeting discussing among other things the problem/solution approach with Board of Appeal legal members; the annual Venice judges’ conference is at the end of next week. All this means that European patent judges have a good idea of what is going on across the system, and an appreciation of one another’s case law that is much richer than that which could be had just from reading cases. I think it is quite unlikely that on a significant point of European patent law an English judge would be wholly unfamiliar with an important development.

Third, the system is underpinned by the high standards among litigators here. It would obviously be wrong selectively to withhold relevant adverse law from another European jurisdiction, just as it would with UK authorities.

How, then, will the UPC affect the exercise of seeking consistency? A recent positive sign is the decision that all judgments will be made public (with an exception for commercially confidential material). It may seem odd to an outsider that this was ever in doubt, but it is clearly a good thing that they will be.

But that’s a point of detail; what we really will want to know is what substantive approach is taken on the big questions. On that, it is too early really to be able to tell. Much will depend on how the UPC organizes itself. Just as no single decision of an individual foreign court could really be decisive in its effect on our law of patent infringement, no single decision of a first instance UPC court is likely to be, and nor, probably, will a single appellate decision be, unless it is clear that it is an authoritative statement of general application.

We should also think about this from the perspective of the UPC itself and of those jurisdictions that have joined it. Those jurisdictions don’t have identical views to each other on all issues, as I have discussed. Will they treat the UPC as a commodore? That question is important at both a theoretical and a practical level.

In any event, in due course there will be judgments from the UPC’s appellate courts that they regard as authoritative guidance on the big issues. It may be relatively soon: its senior judges will be highly expert and very motivated. They will appreciate the importance of a consistent approach in their system, and the benefit of providing it sooner if possible.

Those judgments might be specific and they might be more general. For example, a judgment that claim scope has to be assessed by reference to the Protocol but that no particular methodology is mandatory would be something that every state could live with and needn’t have much impact on us. One would have to wait to see how things worked out over a longer time frame with more cases. On the other hand, to take a hypothetical example, if it were decided that UPC courts will use Improver questions on the same lines as Germany or Spain that might be seen as inconsistent with our approach.

When a consensus on some aspect of infringement emerges in the UPC system that is different from our existing law, the question will present itself of how we deal with that. If it were only subtly different, or if it were not so much different as filling in a gap that isn’t already covered here, then there would still be a good case for acting as in Grimme and taking account of it, preferring consistency wherever possible. The fact that such a consensus emerged from a system in which we are not participating wouldn’t seem to be a reason for downgrading it in itself; it might make very good sense. And it could well come from individual judges whose decisions we have already shown great respect for when they were national judges in their respective jurisdictions.

But for all that, if such a consensus were truly and significantly different from a course we were already set upon by binding authority we would have to decide what to do. The Actavis v Merck exception as decided would seem not to apply – that is specific to the EPC, to the commodore, and clearly it would be significant in that kind of situation that we had chosen not to participate in the UPC system; the role of the Court of Justice could be another factor.

So there would have to be a decision about whether to extend Actavis v Merck. If it were extended then the Court of Appeal could change the law even in the face of its own earlier decisions. Otherwise, it will be a matter for the Supreme Court.

I have struggled to come up with a suitably nautical analogy for the UPC situation. Two convoys? Can you have a convoy of one ship? Probably not two convoys with one commodore but two convoys sharing a GPS signal? Or meeting at sea from time to time? I hope we won’t find ourselves entirely at sea!

Lacking an analogy I nonetheless feel able to say that these are deep waters. As I said at the beginning of this lecture, these matters, and now I am grossly mixing my metaphors, are above my pay grade. So I will leave it there before the waves close over my periscope. Thank you for your attention.

  • 1

    Grimme Maschinenfabrik v Scott [2010] EWCA Civ 1110.

  • 2

    [2017] UKSC 48.

  • 3

    Warner-Lambert v Generics (UK) [2018] UKSC 56.

  • 4

    Subject to the Court of Appeal finding indirect infringement in a respect not relevant to this discussion.

  • 5

    Kirin-Amgen v Hoechst Marion Roussel [2004] UKHL 46.

  • 6

    [2008] EWCA Civ 444.

  • 7

    [1991] RPC 305.

  • 8

    [2011] UKSC 51.

  • 9

    [2019] UKSC 15.

  • 10

    [2007] EWHC 1040 (Pat).

  • 11

    Biogen v Medeva [1996] UKHL 18.

  • 12

    [2008] EWCA Civ 311. In due course the Supreme Court upheld the Court of Appeal decision, [2009] UKHL 12.

  • 13

    [1994] OJ EPO 695.

  • 14

    My thanks to Nigel Stoate and Dieter Kamlach at Taylor Wessing.

  • 15

    Improver Corporation v Remington Consumer Products [1990] F.S.R. 181.

  • 16

    [2013] UKSC 46.

  • 17

    IPCom v HTC [2013] EWCA Civ 1496.

  • 18

    Fujifilm v Abbvie [2017] EWCA Civ 1.

  • 19

    R v Secretary of State for Work and Pensions [2008] UKHL 63.

  • 20

    At [67].

  • 21

    And thanks to Richard Arnold for giving me a copy.

Contributor Notes

This lecture was given before the Enlarged Board of the EPO gave its decision in G2/21 and before the UPC Agreement entered into force on 1 June 2023

Since 2022 Since May 2022 Past 30 Days
Abstract Views 249 249 0
Full Text Views 676 676 39
PDF Downloads 307 307 28