Improving Appeal Conferences At The USPTO | December 2012

By: Joseph J. Mallon, Knobbe Martens Olson & Bear LLP, and Kate S. Gaudry, Kilpatrick Townsend & Stockton LLP
Publication: Law360
Download PDF

Law360, New York (December 03, 2012, 12:16 PM ET) — Patents can provide incentives for entities to invent and to develop and market ideas. However, to achieve this outcome, examination by the United States Patent and Trademark Office must be structured such that deserving applications are routinely allowed to issue. Assessing such patent-examination quality in an objective manner is difficult. The USPTO currently uses a metric developed in concert with the Patent Public Advisory Committee that focuses on quality of issued actions, use of best practices, efficiency evaluations, and survey-based perceptions of applicants, practitioners and examiners.[1]

The quality of issued actions is evaluated using a sampling-based approach to assess the presence or absence of rejections using a reasonableness or clear-error standard. This assessment does not include a final determination about whether rejections are legally or factually valid, and thus, whether an applicant is being unfairly denied patent protection.

Decisions by the Patent Trial and Appeal Board do address this concern, as an independent panel of judges substantively determines whether rejections should be affirmed or reversed. For fiscal year 2012, the USPTO reports an overall total affirmance rate of 49.3 percent.[2] However, the value of the overall affirmance rate as a measure of examination quality is questionable, as the cases decided by the PTAB are not a random sampling of rejected patent applications.

Rather, they have been subject to a selection process by both applicants and the USPTO, often in a nonsymmetrical and idiosyncratic manner. In deciding to initiate and continue an appeal, applicants typically consider a number of factors, including availability of alternatives (such as abandonment or request-for-examination filing), potential claim scope, potential for patent-term adjustment (PTA), cost of appeal, delay resulting from appeal process, uncertainty of appeal, etc.

For example, for an application covering an important pharmaceutical, an applicant might reasonably choose to appeal even a relatively solid rejection because amendments to reduce claim scope would render the resulting patent much less valuable and the potential for significant PTA and desirable claim scope would justify the cost, delay and uncertainty of the appeal process.

On the other hand, the decision by the USPTO to continue an appeal would appear to be much more objective. Internally, the USPTO conducts appeal conferences to review appeal briefs submitted by applicants. The conferees are typically the examiner, a supervisory patent examiner (SPE), and another SPE (or in some cases a quality assurance specialist). If the conferees agree with the examiner’s rejections, they signify their approval by signing the examiner’s answer and the case proceeds to the PTAB for decision.

If they do not, the examiner is directed to reconsider them, typically resulting in allowance or reopening of prosecution with new grounds of rejection. The decision by the conferees would appear to be driven primarily by their own collective determination of the merits of the rejection, and not by more idiosyncratic applicant-specific factors such as the economic value of claim scope, PTA, appeal cost and delay.

The relative objectivity of the USPTO and the involvement of supervisory personnel at the appeal conference stage would be expected to increase the likelihood that appeals reaching the PTAB would involve high-quality rejections and would typically be affirmed. However, achieving this result depends on the degree to which supervisory personnel provide quality review of pending rejections.

To address this question, we analyzed over 2,700 PTAB decisions on appeals from Technical Center (TC) 1600 published since 2007. Twenty-six SPEs participated in appeal conferences for upwards of 50 decisions (averaging over 100) and thus, were suitable for meaningful statistical review. For each such SPE, we calculated how often the PTAB affirmed the rejection that the SPE had supported at the appeal conference stage, as indicated by signing the examiner’s answer.

The resulting SPE-by-SPE affirmance rates are provided in the table below. Although the names of the individual SPEs are available in PAIR and thus a matter of public record, we have elected to refer to them as SPE A, B, C, etc. For comparison, the mean average affirmance rate for all PTAB decisions in TC 1600 over this time period (47 percent) is also included in the table.

The data shows that the affirmance rates differed dramatically depending on which SPE reviewed the matter: the lowest SPE-specific affirmance rate (29.6 percent) was approximately half that of the highest affirmance rate (58.3 percent). We conducted a shuffling analysis to determine how likely the observed rate discrepancies resulted from random sampling of a distribution. Specifically, for the decisions analyzed for the 26 SPEs, we identified how many affirmances and how many total decisions were associated with each SPE.

We then randomly shuffled the affirmances across SPEs 10,000 times. The probability of achieving a shuffled data set with inter-SPE variability greater than that actually observed (actual variance of 54.8 percent) was 0.32 percent. The probability of achieving an affirmance rate as low or lower than the lowest SPE-specific affirmance rate was 3.2 percent. Thus, our data indicates that the inter-SPE variability is substantial and statistically significant.

From a human factors perspective, one does not really expect statistically equivalent performance from 26 different SPEs having various skill and experience levels. However, the USPTO could at least attempt to address this variability to improve the probability that an applicant receives a prompt and reasonable evaluation of pending rejections without needing to undergo the time and expense of the full appeal process.[3] Specifically, USPTO management could identify outliers and implement appropriate corrective measures.

For example, the USPTO could collect appeal data such as that shown in the table and use it to redesign appeal conferences to include only those SPEs having a better than coin-flip ability to be affirmed by the board. The remaining SPEs could be retrained or reassigned to tasks more suitable to their skills. This approach would enhance overall examination quality by reducing the number of deficient rejections going up to the PTAB, reduce the PTAB backlog, and provide more accurate feedback to the examiner regarding the shortcomings of the rejections.As a corollary, we also suggest that the USPTO consider implementing a “fast-track” appeal procedure, similar to the current Track 1 procedure for patent examination. Such a Track 1 appeal would provide many of the same benefits as Track 1 examination, and also improve examination quality by providing faster PTAB feedback to the SPE and examiner about the appropriateness of rejections at issue.

In conclusion, the appeal process provides applicants with an opportunity to have erroneous rejections withdrawn, such that worthy inventions can be patented. Additionally, the USPTO’s procedure of having SPEs review rejections prior to submitting the matter to the PTAB provides an opportunity to reduce the probability that applicants need to resign themselves to the temporal and financial disadvantages associated with appeals. However, this advantage will only be achieved if the review by these SPEs is high quality.

Our data shows that the quality of the review varies dramatically across SPEs, with some SPEs being associated with surprisingly low affirmance rates. Initiatives by the USPTO to address at least these low outliers would improve the overall operation of patent examination and improve the patent system’s ability to address its innovation-encouraging objective.

Table: Each row is associated with a SPE in TC 1600 who participated in 50 or more appeals. The second column shows the percentage of such instances in which the PTAB subsequently affirmed all pending rejections.

SPE All-Affirmance Rate (%)