How to instruct experts in patent cases correctly

Written on 18 Dec 2020

The recent decision of Mr Justice Meade in Fisher & Paykel v Flexicare is yet another warning to patent practitioners of the dangers in failing to instruct experts properly, particularly when the expert has prior knowledge of the invention.

Every patent practitioner will appreciate the importance of properly instructing expert witnesses and the potential consequences of not doing so correctly.

The 'Medimmune' guidance

Experts involved in patent cases often have a closer degree of collaboration and interaction with instructing lawyers compared to experts involved in other commercial disputes.  This is because they need to familiarise themselves with various patent law concepts in order to ensure their evidence assists the Court, whilst the legal team needs to learn the technology in detail to understand properly what the expert is saying.

Through this (often lengthy) process of taking evidence, it is critical that experts remain impartial and independent, and are able to express objective unbiased opinions on matters within their expertise.  The lawyers instructing them bear a heavy responsibility to ensure that experts maintain their objectivity throughout the process of taking evidence.  Failure to do so is likely to become apparent during cross-examination at trial and completely undermine the expert's credibility.  The result is that the trial judge will be unwilling to place much if any weight on that expert's evidence, which may be determinative of key issues in the case (e.g. obviousness and sufficiency).

This is the context in which Mr Justice Arnold (as he then was) provided influential guidance in Medimmune v Novartis [2011] EWHC 1669 (Pat).  It is now common practice to provide experts with key extracts from this judgment at the outset of their engagement.

In that case, Mr Justice Arnold identified two common problems encountered in failing to properly instruct experts in patent cases ("traps for the unwary" as he put it).

  • The first is failing to give a balanced view on the disclosure of a particular item of prior art, for example by failing to discuss unclear or potentially unhelpful passages.
  • The second occurs when an expert was previously aware of the invention in issue, or had some involvement with it or a similar invention (for example, because they published relevant papers or are listed as an inventor on a patent application in the same field), but has failed to explain this information adequately (or at all) in their reports.

Concerns over loss of objectivity may also arise if an expert fails to refer to key papers or developments in the field.  A recent example of the problems arising out of this sort of approach can be seen in Neurim v Generics [2020] EWHC 3270 (Pat).  Here, the written evidence of the defendant's expert was heavily criticised by the judge as being disingenuous and "calculated, not to assist, but to mislead, the court".  The expert failed to refer to key papers of which he was previously aware and he was also, in the judge's analysis, guilty of misleading citation of material and misdirection. The result was that the judge was unable to accept his evidence on certain key points.

Medimmune v Novartis also approved the approach of showing documents to experts in a chronological sequence. This involves first discussing the skilled person/team, then the common general knowledge, followed by the prior art and, finally, any patent documents.  This process is sometimes known as 'sequential unmasking' and is a method of trying to minimise or avoid an expert expressing opinions with a hindsight bias (the hindsight bias being knowledge of the patented invention).  This is especially important for parties seeking to revoke a patent on the grounds of obviousness.

The problem with hindsight

Hindsight is inherent in any system that must look back in time to analyse the validity of a patent.  Patent cases commonly involve inventions conceived of many years before proceedings are issued.  The parties advance positions about what the state of the field was and how particular problems would have been approached at a particular point in time.  This is necessarily a hindsight exercise.  Yet, the legal approach to assessing whether an invention was obvious at a particular point in time requires the experts artificially to put out of their minds any knowledge of the invention and subsequent advances in the field.

Hindsight can manifest itself in a number of ways when it comes to an expert's consideration of what would have been obvious to do at a particular point in time (aside from the classic "step-by-step"/Technograph problem, a number of others are identified by Lord Justice Floyd in E. Mishan & Sons v Hozelock [2020] EWCA Civ 871 at [98]).  Many brilliant inventions could be considered to be 'simple' in the sense that they seem to be obvious after they have been made.  These inventions may not have required a particularly complex intellectual leap but rather simply conceiving of an idea in a (non-obvious) way that had not been done previously.

Hindsight is a dangerous thing indeed and a failure to properly identify it and exclude it would impose an unrealistically high – even impossible – threshold for patentees to overcome.  The courts are well-aware of the dangers of hindsight reasoning and the need to exclude it (see Actavis v ICOS [2019] UKSC 15 at [72]). It should be in all parties' interests to try to exclude hindsight reasoning from their expert evidence and arguments at trial as a failure to do so can be fatal to a party's case.

The need to exclude hindsight knowledge or bias is also accepted as being of fundamental importance in other fields (outside the somewhat artificial exercise conducted in patent cases). Most notably, the use of blinding (or masking) in the design of randomised controlled trials for the development of new medicines or developments in clinical practice.  There is evidence showing that clinicians find what they expect and hope to find so blinding both the patients receiving the treatment (or placebo) and the clinicians administering the treatments (so-called 'double blinding') is a way of eliminating or minimising this performance bias.  The use of blinding is accepted to provide better quality evidence for clinical assessment, and double blinded trials carry more weight than unblinded or single-blinded trials.

The blinding of experts (by not telling them which party they are instructed by and/or who is the patentee) may help minimise any loss of objectivity but it does not really address the issue of hindsight bias in patent cases.

Practical problems with sequential unmasking

Earlier this year, in Akebia & Otsuka v Fibrogen [2020] EWHC 866 (Pat), Lord Justice Arnold (sitting as a judge in the Patents Court) felt the need to re-iterate the Medimmune guidance as he didn’t feel it was being sufficiently well-observed by patent practitioners.

The Fibrogen case raised a number of issues concerning the failure to properly instruct experts. One of which was a failure by one of the experts to refer to a number of papers on which he was a co-author and which were highly relevant to the field of the invention (the second trap for the unwary dealt with in Medimmune). The judge said this gave him cause for concern over the preparation of the expert's reports, commenting that he should have been instructed to address these papers in his reports.

Lord Justice Arnold re-iterated his approval of the sequential unmasking protocol explaining that it can help to reduce or avoid hindsight and this is desirable wherever possible.  But he also pointed out that it is not a rule or principle which must be followed in all circumstances. He identified three practical problems associated with following the protocol, which are:

  • The issue of identifying the skilled person/team before seeing the patent and discussing the common general knowledge on this basis. Patents are addressed to those individuals with a practical interest in putting the invention into effect so the skilled person/team can only be properly identified and approved by the expert after having seen the patent. In practice, this is done by the lawyers asking the expert to adopt their characterisation of the skilled team  pending a review of the patent.  Lord Justice Arnold considered that this practice is imperfect but it is necessary in order to be able to follow the sequential unmasking process.
  • When the party seeking revocation of the patent amends its case to introduce new prior art after the expert has reviewed the patent. This is what happened in Fibrogen and is a risk in any UK patent proceedings.  In these circumstances, the expert must do their best to avoid hindsight reasoning and explain how they have sought to do so.  Lord Justice Arnold did not criticise the way in which the expert was instructed in this instance, but he did take it into account when considering obviousness.  This was not an issue that ultimately undermined that expert's evidence on obviousness.
  • Pre-existing knowledge of the expert. In this case, the expert was aware of the class of inhibitors claimed by the patent.  Lord Justice Arnold accepted that, as a result, it would not have been possible to instruct the expert in a way which was free from hindsight.  This is the issue which is addressed more directly in the recent Fisher & Paykel case (below).

Those in favour of sequential unmasking

Some judges of the Patents Court are particularly in favour of sequential unmasking as a way of avoiding or reducing hindsight bias (e.g. Lord Justice Arnold); others are less in favour of it (for example, Mr Justice Birss; see HTC v Gemalto [2013] EWHC 1876 (Pat)).  The sequential approach is often criticised by practitioners (and clients) due to the length of time taken (and costs incurred) to reach a point with an expert which allows the legal team to assess their suitability, particularly given the time frames available to respond to claims for patent infringement.

While there is no formal rule requiring this approach, the risk that the ultimate trial judge finds that the expert was instructed incorrectly may undermine the weight the judge attaches to the evidence.  This can have very significant consequences on the outcome of the case. It is rarely worth the risk of not following this prudent sequential approach to the instruction of experts.

Mr Justice Meade in favour of sequential unmasking?

With the recent appointment of Mr Justice Meade to the Patents Court, patent practitioners will be curious as to which camp he falls into (after all, he was lead counsel for Medimmune in the case which provided the guidance discussed above).  The decision in Fisher & Paykel v Flexicare makes clear that, whilst acknowledging it may not always be practical, Mr Justice Meade does indeed see value in this approach ([19] to [22]), even if the expert is already aware of the patent in issue or a commercial embodiment of it.

Importantly, in such circumstances, Mr Justice Meade explained that experts cannot give completely untainted views of the prior art but may nevertheless, with extra care and consideration, give cogent evidence on obviousness avoiding hindsight bias.  He explained (in [21]) that "in such a situation I think it must be important for the expert to identify how they knew about the invention and when, and to reflect carefully on how that might influence them."

Fisher & Paykel v Flexicare

In Fisher & Paykel, the defendant's expert saw the patent some years before his involvement in the litigation. He was also aware of the features of the claimant's commercial products which were relevant to the patent (including having referenced them in his own patent application).  Although his first report contained a "mere reference" to having previously seen the patent, the judge considered that this was "inadequate because he had not thought about his previous knowledge of the alleged invention properly, and this lack of care meant that he had not really turned his mind to how it might have influenced his thinking on the issue of obviousness".  Ultimately, the expert's views were influenced by "significant hindsight" and he failed to articulate proper reasons to support his conclusions that each prior art citation rendered the patent obvious.  The obviousness attack failed and the patent was upheld as valid (and infringed).

The important point from this decision is that if an expert is already aware of the patented invention before their involvement in proceedings, they must thoroughly explain the circumstances in which they became aware of the invention, give proper consideration as to how this may influence their assessment of the prior art, and also explain how they have sought to avoid hindsight reasoning.

Following the sequential unmasking protocol and getting the expert to reach conclusions at each stage of the process may help the expert to anchor their thinking at particular points in time to minimise hindsight bias.  Asking an expert whether there were particularly significant work or life events around the priority date can sometimes be helpful, but there is rarely an easy answer.

Experts are usually being asked to consider events which took place many years in the past, when they had little reason to remember contemporaneously exactly how and when developments were being made.  Practitioners should therefore always give careful consideration at the outset of a case as to the order in which documents should be shown to potential experts, and take great care to manage any hindsight reasoning (whilst, of course, ensuring that the expert remains objective throughout).