Reading PTAB Trial Outcomes: Institution, Settlement and Final Decisions

Numbers about the Patent Trial and Appeal Board are quoted constantly: institution rates, settlement rates, the share of final written decisions that cancel every challenged claim. They are useful, but only if you know what is being counted. A single trial can end in several different ways, and a single patent may be challenged in several petitions.

This page does not report current figures. For those, see PTAB statistics. Instead, it explains the possible endpoints of an AIA trial and how outcome rates are built, so that any statistic can be read with care.

The stages where a trial can end

An inter partes review or post-grant review moves through a predictable sequence: petition, optional patent owner preliminary response, institution decision, trial, oral hearing and final written decision. A proceeding can stop at almost any point.

Denial of institution

The Board, acting for the Director, decides whether to institute. For IPR the petitioner must show a reasonable likelihood of prevailing on at least one challenged claim (35 U.S.C. § 314(a)); for PGR the standard is whether it is more likely than not that at least one claim is unpatentable. See the different threshold standards.

A petition can be denied on the merits, because the threshold is not met, or on discretionary grounds. A denial is generally not appealable (§ 314(d)).

Settlement and termination under § 317

The parties may settle at any stage. Under 35 U.S.C. § 317(a), an IPR is terminated as to a petitioner on the joint request of the petitioner and patent owner, unless the Office has already decided the merits before the request is filed. If no petitioner remains, the Board may terminate the proceeding or proceed to a final written decision. Settlement agreements must be in writing and filed with the Office (§ 317(b)), and a party may ask that the agreement be kept separate as business confidential information. Section 327 provides parallel rules for PGR.

Adverse judgment

A party may request adverse judgment against itself, and certain actions are treated as such a request under 37 C.F.R. § 42.73(b). Examples include a patent owner disclaiming all challenged claims, cancelling claims so that none remain in the trial, or abandoning the contest. The result ends the trial against that party.

Final written decision

If a trial is instituted and not otherwise terminated, the Board issues a final written decision on the patentability of every claim challenged in the petition and any new claim added by amendment (35 U.S.C. § 318(a)). Since SAS Institute Inc. v. Iancu (2018), institution is all-or-nothing as to the challenged claims. A final written decision may hold:

• All challenged claims unpatentable
• Some claims unpatentable and others not shown unpatentable
• No challenged claim shown unpatentable

After appeals end or the time for appeal expires, the Director issues a certificate canceling claims found unpatentable and confirming those found patentable (§ 318(b)).

How outcome rates are computed

The same docket can produce very different headline numbers depending on three choices.

Choice	Options	Effect on the reported rate
Unit of counting	Petition, patent or claim	A patent challenged in three petitions counts once by patent, three times by petition
Denominator	All petitions filed; petitions reaching an institution decision; instituted trials; final written decisions	Excluding settled or pending cases raises or lowers rates sharply
Time frame	Fiscal year of filing, of decision, or cumulative since 2012	Policy changes over time affect older and newer cohorts differently

Counting by petition, patent or claim

Petition-level rates are the most common, but they can mislead when one patent faces several petitions, some of them denied as redundant. Patent-level rates ask whether a patent was challenged successfully in any proceeding. Claim-level rates ask how many individual claims were held unpatentable, and often differ greatly from petition-level figures because a petition may challenge many claims but succeed on only a few, or the reverse.

Choosing the denominator

An “all claims unpatentable” rate looks very different when computed against final written decisions alone than when computed against all petitions filed. Many petitions settle before or after institution, and some are denied. Whether those cases are included is often the biggest single reason two published rates disagree.

Discretionary denials and institution rates

Institution rates also reflect policy, not only the strength of petitions. The Board may deny institution on discretionary grounds, for example under § 314(a) in light of parallel litigation or serial petitions, or under § 325(d) where the same or substantially the same art or arguments were previously presented to the Office. The Office’s approach to discretionary denial has changed several times, so institution rates from different periods may not be comparable. Check current USPTO guidance before drawing conclusions about a particular period.

Reading statistics carefully

When evaluating a statistic, ask:

1. What is being counted: petitions, patents or claims?
2. What is in the denominator, and are settled or pending matters included?
3. What period does it cover, and did relevant policy change during it?
4. Does it combine IPR, PGR and the former CBM program, or separate them?

For the overall process, see the inter partes review guide and the Patent Trial and Appeal Board guide.

Key takeaways

• A trial can end by denial of institution, settlement under § 317, adverse judgment under § 42.73(b), or a final written decision.
• Final written decisions can find all, some or none of the challenged claims unpatentable.
• Reported rates vary because they count by petition, patent or claim and use different denominators.
• Discretionary denial policy affects institution rates and has changed over time.
• For current figures, see PTAB statistics and USPTO publications.