02 December 2012

Study Shows that Patent Examiners Make a ‘Meaningful Difference’

AU-EP-US (Images by Geoscience Australia, www.ga.gov.au) On 21 November 2012, we attended a seminar in Melbourne presented by Professor Andrew Christie of the Intellectual Property Research Institute of Australia (IPRIA) on the topic of An Empirical Comparison of the Outcome of Patent Examination in the USPTO, the EPO and IP Australia.

While the title may seem dry, Professor Christie provided, as always, an intelligent, engaging and thought provoking presentation (for an earlier example, see Why IP Professionals Must Take ‘Gene Patent’ Opponents Seriously).

IPRIA researchers studied a sample of 494 patent families, on which patents had been granted by all three of the US, European and Australian patent offices in recent years, and for which the main claim of each application was identical in all three offices when originally-filed.  They compared the granted claims with the original claims in each case, to identify whether the examination process had resulted in a ‘meaningful change’ to the scope of the claims.

What the researchers found was that in a majority of cases, across all three patent offices, examination had resulted in the applicant amending its claims in some meaningful way.  This happened most commonly in the USPTO (79% of cases), while the rate of meaningful amendment in the EPO was 68%, and in Australia 57%.  Most of the amendments were what the researchers called ‘integral’ changes, i.e. the narrowing of the main claim by incorporation of additional features or limitations.

Somewhat surprisingly – despite a common perception that broad ambit-claiming is more common in some fields of technology that others – the study identified no significant difference in the rate of amendment across different technical fields.

In a second phase, the researchers took a subset of 116 applications, and studied the records of examination in all three offices to identify the causes of amendment, and the differences between the final granted claims between the three offices.  What they found was that prior art grounds (i.e. novelty and inventive step rejections) were overwhelmingly the cause of amendments.

Furthermore, they found that in a majority of cases the granted claims were meaningfully different between at least two of the three offices, with the USPTO tending to restrict the scope of claims to a greater degree than the EPO.  The Australian Patent Office was most likely to grant broader claims.


For the purposes of the IPRIA study, the assessment of whether or not an amendment was ‘meaningful’ was made by experienced patent specialists.  The researchers essentially had to read and compare three versions of claim 1 for each of the 494 cases in the study.

Differences which resulted solely from different formal practices in each jurisdiction, and which did not alter the actual scope of the claims (such as the common replacement of the word ‘comprises’ with ‘includes’ in Australia) were considered not to be meaningful.  However, any amendment which resulted in the broadest claim having a different scope at grant than at filing was considered a ‘meaningful change’.

In the first phase of the study, examination was treated as a ‘black box’, i.e. the input was the original claim, and the output the granted claim, and no information was obtained about what went on the examination process.

As noted above, meaningful change occurred in 79% of cases before the USPTO, 69% of cases in the EPO, and 57% of cases at the Australian Patent Office.  The different outcomes were significant, both in practical terms (i.e. patents with differing scope) and in the statistical sense.  The influence of the Office on the outcome was shown to be a statistically-significant factor, with a p-value less than 0.001 (i.e. there is less than a one in one-thousand chance that the differences are purely random variation resulting from the choice of the sample set).

Patent examiners will therefore be pleased to know that the really do make a difference!

Interestingly, the other factor indentified by the researchers which was of a similar level of significance was country of origin of the original applications.  Swedish applicants, for example, were far less likely to amend before the EPO (50% of cases) than Australian applicants (91% of cases).  However, only 55% of those same Australian applicants amended their Australian applications, compared with 38% of Swedish applicants.


The second phase of the study looked at the rate of meaningful differences between patent offices.  The definition of ‘meaningful’ was the same as for ‘meaningful change’, i.e. that the difference in claim wording corresponds with a real difference in the scope of the patent.

Unsurprisingly, prior art did most of the ‘heavy lifting’ in causing meaningful differences in examination outcomes, with objections of lack of novelty and/or lack of inventive step (i.e. obviousness) being the main cause of claim narrowing.  The end result of this was that claims were more likely to be narrowed in the US, followed by Europe, with the Australian Patent Office tending to grant the broadest claims.

Interestingly, while one might expect that the examiners in all three offices would rely on the same prior art – particularly in view of the applicant’s duty of disclosure in the US, and the availability of the US and European prosecution histories online throughout the period of the study – this was not necessarily so.  In 78% of cases in which claims were narrowed, European and US examiners cited different prior art when objecting to claims, while in about two-thirds of cases Australian and European examiners relied on different prior art.

As between US and Australia, USPTO examiners cited the same prior art as their Australian counterparts in around half the cases in which the US claims were narrower.  There were only two cases in which Australian granted claims were narrower than their US equivalents, and in both of these cases the Australian examiner cited only prior art that had also been raised by the USPTO.


Firstly, we should point out what the result of this study do not tell us.  They say nothing about ‘quality’ of examination.  While quality is difficult to assess in practice (despite many measures being proposed), the only true definition of ‘perfect’ examination is where the claims granted are no broader, but also no narrower, than is justified in view of the prior art (i.e. the inventor is given rights precisely coincident with the contribution made by the invention to the stock of human ingenuity). 

While there is greater public concern about the grant of unduly broad patents, the system fails equally when it denies rights to an applicant that should be available under the law, and some examiners are certainly tenacious in maintaining rejections of dubious merit.  This may be especially true in the US, where the law can be fairly liberal regarding combination of prior art documents in an obviousness rejection.  Many would argue that if it is necessary to combine five or more documents to cover all of the features of a claim, this could not generally be said to be ‘obvious’.  Yet this is not an uncommon scenario.

So, differences in claim scope – either broadening or narrowing – tell us nothing about quality.

They do, however, tell us something about differences in law and examination practice.  While further study may be required to draw any firm conclusions about the origins of amendments, and difference in outcomes, there are a few observations that might be made.

Firstly, the fact that patents granted in the US tend to be narrower in scope than in Europe or Australia is, on reflection, unsurprising.  With regard to the US situation, the scope of prior art for the purpose of assessing obviousness is greater in the US than in other jurisdictions, because it includes unpublished US patent applications with an earlier filing date.  In Europe and Australia, such prior art is available only for assessing novelty.  Furthermore, the decision of the US Supreme Court in KSR v Teleflex provides a very liberal standard for examiners to find claims obvious.  Furthermore, the ‘broadest reasonable interpretation’ standard applied to US claims may result in applicants using narrower language in order to overcome prior art rejections than they might require in Europe or Australia.

In Australia, by contrast, the law in-force during the period of the study places explicit limitations on the application and combination of prior art documents for the purpose of assessing whether or not a claim involves an inventive step.  This is soon to change, under the Raising the Bar reforms, and we wrote about the problems with the current provisions in Patent Reform Exposed, Part I – Raising the Inventive Step.  However, for all patents and applications not covered by the revised laws, it should be substantially more difficult to establish a lack of inventive step in Australia, compared with the same claims in the US or Europe.

This probably explains why meaningful change occurred less frequently in Australia than in Europe and the US, and also why Australian granted claims were often broader than their US or European counterparts.  It is also consistent with another result of the study – although claims were less-often amended in Australia, when there was meaningful change to claim scope it was due to novelty objections in a massive 85% of cases, compared with only 69% and 66% respectively in the US and Europe.  This suggests that Australian examiners rely more heavily on lack of novelty as a valid objection to claims in Australia, although curiously the ratio of novelty to obviousness objections appears to have been very similar across all three offices.

It is unsurprising to find that the EPO lies somewhere between Australia and the US in terms of claim-narrowing.  The prior art available in Europe and Australia is similar in content, however European examiners are not so restricted as Australian examiners in applying the prior art for the purpose of assessing inventive step.  Even so, it is our experience that European examiners are less inclined than their USPTO counterparts to raise objections based on combinations of more than two or three prior art documents.


While examining for novelty and inventive step (i.e. nonobviousness) is the most important role of the patent offices, there are other objections that can cause meaningful changes in claim scope.

One of these is lack of clarity, generally meaning that the language of the claims is somehow ambiguous, such that the precise scope cannot easily be ascertained.  The survey confirmed what all experienced practitioners know – that the EPO has far stricter standards of clarity than either the USPTO or Australian Patent Office.  Of all the cases in which meaningful amendments were made, this was due at least in part to clarity objections on 26% and 28% of occasions in the US and Australia, respectively.  The rate of amendment due to clarity issues in Europe was much higher, at 56%!

On the other hand, the requirement for a sufficient ‘written description’ of the claimed invention (or its equivalents in Europe and Australia) contributed to roughly the same proportion of meaningful amendments in all three patent offices (7-8%).  This no doubt reflects that fact that, despite some differences in expression of the legal requirements, the substantive tests in each jurisdiction are not so very different.

In a few cases, meaningful changes also occurred due to objections of lack of unity (i.e. claims directed to more than one invention).  In Australia and Europe, where the law on unity of invention is essentially the same (and equivalent to the corresponding PCT requirement), the rate at which this ground contributed to amendment was roughly equal (3% and 4% respectively).  However, US examiners’ requirements for Restriction/Election (the US unity equivalent) affected 10% of amended cases.

While the US national law on unity of invention is quite different to other countries’, the disparity in the case of IPRIA’s sample is of concern.  One of the selection criteria for the 116 application families used in the second phase of the study was that they all originated from common PCT applications.  When considering applications that arise as the US national stage of a PCT application, USPTO examiners are supposed to apply the PCT unity requirement, rather than follow US restriction practice (37 CFR 1.499).  So for this sample, there should have been no difference in the unity standards applied in the three patent offices.  However, practitioners accustomed to prosecuting applications out of the PCT across various national patent offices will probably not be surprised to find US examiners raising unity objections at more than twice the rate of the European and Australian offices.  This is certainly consistent with our experience, and generally seems to result from the relative lack of familiarity of US examiners in applying the PCT rules.


Overall, IPRIA has produced an interesting study, which confirms some anecdotal beliefs about the comparative outcomes of examination in different patent offices, as well as providing a few useful – and sometimes even surprising – insights.

The main conclusions of the work, as identified in Professor Christie’s presentation, are:
  1. the most common ground causing meaningful amendments to claims is lack of novelty (suggesting that applicants commonly err on the side of reserving greater rights, by initially claiming too broadly);
  2. in the US, narrow claims result from greater application of prior art, while in Europe this is more commonly the result of a more stringent application of legal tests;
  3. patent examination has a practical effect on outcomes in all three offices, i.e. patent examiners make a real difference;
  4. the magnitude of this effect, however, varies significantly, with US examiners ‘causing’ the greatest meaningful change in scope, and Australian examiners the least.
While the study raises many new questions, it seems that we cannot expect any further data-gathering at this stage, because the funding has run out.  However, it is likely that we will see a number of papers generated, as the existing data is mined for new insights.

No formal peer-reviewed papers have been published as yet, however the slides from Professor Christie’s presentation – including a few handy charts and tables – are available from the IPRIA website.

No comments:

Post a Comment