Citing References at the PTO

So far in 2012, the USPTO has issued just about 200,000 utility patents. As part of the application process, patent applicants have a duty to identify prior art references that are material to patentability. About 13% of applicants submit no prior art. The median submission is 8 references. However, the data is highly skewed. While most entities submit a relatively small number of references, there are a few entities that submit a large number of references for consideration at the USPTO. Amazingly, the top ten percent of patents (in terms of references cited) are associated with over two-thirds of the seven million prior art citations in this set. Although the median number of references cited by applicants is 8, this skew shifts the average way up to 36 references. The submission of four or five references can be quite helpful. The submission of forty references is unlikely to be helpful (absent a second level search tool). In the past, this has prompted the USPTO to consider mechanisms to place additional burden on applicants who feel the need to submit such a large number of documents for consideration.

Citation also varies by country of origin. The chart below shows average applicant citation count according to assignee home country.

28 thoughts on “Citing References at the PTO

  1. 28

    But cheer up, you can always circle your friends together to ignore reality and talk to each other without any opposing views.

    An even safer approach, IANAE, is to do what anon was caught doing: invent 30 friends and then circle them together to ignore reality and talk to each other without any opposing views.

  2. 27

    I can see that facing reality (especially given your predilections) is not fun. But cheer up, you can always circle your friends together to ignore reality and talk to each other without any opposing views.

  3. 26


    Yeah. It’s depressing, but at least it’s trained me to not worry about convincing anon. I’m satisfied if I make my point clear to the sane readers like you and Max who happen by from time to time. It’s made the internet much more fun.

  4. 24

    Not only too subtle – but goes against the very grain that IANAE has long preached upon: the right is not the remedy.

    Guys – if you want to be pedantic and take shots, at least practice a little and come close to the target. It’s embarrassing how off the mark you guys are.

  5. 22

    I was asking about requirement – not penalty MaxDrei.

    Don’t tell me it’s possible for the legal requirement and the associated penalty to be different somehow. Got any Supreme Court cases on the point?

  6. 21

    There is the rub MM. Who bears the cost of reveiwing these references? These references can be generated by foreign patent offices, inventors, related cases, etc.

    The fact is that the PTO needs to be the expert. And, when the burden for figuring out references is put squarely where it belongs on the PTO, then all of these “problems” are seen as not problems.

    You say 5 minutes. OK. 200 * 5 at a bargain rate of $350/hour. 1000 minutes. 16.6 hours >$5K.

    And, just think, > $5K for references that may have been generated by others.

    The only way to make sense of this is for the examiner to be an expert in their field.

  7. 19

    ? Not sure how you reached your conclusion Greg.

    How does patent reference data affect patent quality (whatever that is) in the first place?

  8. 18

    At least a third of these “references” are not references in the traditional sense. You need to make a fair number of adjustments to your data to draw any conclusions, starting by defining what you mean by “reference”. Adjusted for non-references and normalized to total potential prior art, patent reference data supports the argument that patent quality has gradually declined for at least 20 years.

  9. 16

    I can tell you how this works in practice for a typical bio application that is filed in the US, PCT and then worldwide. Let’s say the claim is to a Method for achieving X, comprising administering Y and Z, wherein said Z is administered at level W.

    The specification includes reference to, say, 5 earlier protocols for achieving X. Those are submitted. There’s a couple refs that arguably teach away from the invention. So we’re up to 7. Then there’s six references teaching how to make Y and Z, how to use them separately and in combination for a variety other purproses. So we’re up to 13. The case is filed with those references in the US and PCT. Then the ISR/WO comes in from the PCT and it cites four other references. You submit the ISR/WO and the other references. We’re up to 17. After national stage is entered around the world, the substantive exam reports from the foreign jurisdictions come in. You submit those and the accompanying references, say, 8 documents total. Now we’re up to 25. Meanwhile a second related application has been filed and an Office Action has been received in that application, citing three new reference. So we’re up to 30. During prosecution, five patent applications in more or less the same area are published by the client’s competitors and identified by the client as potential relevant. So we’re up to 35.

    That’s a very reasonable scenario, in my opinion.

    I’ll also go out on a limb and say that if you are citing more than 200 references you are probably citing 100 references that could be recognized as cumulative or irrelevant after a five minute review of the file.

  10. 15

    You tell me. What’s the penalty for leaving out something that might turn out to be interesting, and what’s the penalty for including something cumulative? Besides, when two docs are not identical, are they ever “cumulative” of each other?
    Don’t most lawyers err on the safe side. What is the safer of the two sides here?

  11. 13

    Why should in-house cite more? I was simply supposing that it would almost inevitably have a growing document collection and that, with every successive IDS it felt obliged to prepare, it would routinely throw the whole (ever-bigger) collection at the USPTO.

    Tell me won’t you, if my imagination is over-active.

  12. 12

    where do you get this data that is downloadable into Excel (as these appear to be Excel bar graphs and charts)? I’d like to do some of this myself.

  13. 11

    I would have thought the trend might have accelerated because of McKesson (2007), but do not see any particular McKesson related spike. However, I query how many of the references are now McKesson style cross citations of Office Actions and such.

  14. 10

    One reason for citing is that the USPTO published final rules related to Patent Applications Containing Patentably Indistinct Claims in 2007(See, 72 Fed. Reg. 46716). The rules included cross-referencing requirements of applications and patents that are related in filing time, inventorship, and ownership. Furthermore, the rules presumed patentably indistinct claims where such related applications are filed on the same day and contain substantial overlapping disclosure. (see 37 CFR § 1.78(f)). The rules were the subject of litigation, never took effect, and ultimately were removed from the Code of Federal Regulations.
    Nonetheless, prudence dictated citing commonly owned cases that may not otherwise have been cited.

  15. 8


    I accept your states reasons as sound in and of themselves, but still fail to see how this translates to a greater or lower number of items in IDS filings. I could quite easily see that these in-house experts may submit more OR less references in any particular situation. Some type of rationale linking a greater or lower number of citations still needs to be made.

  16. 7

    The “cause or correlation with point 3” is that U.S. in-house patent attorneys [a declining breed] are much more likely to have been filing and prosecuting more applications on the same type of products and technologies [their employer’s], understand the art and technology better, and have a better idea of what is novel and what is not in that technology.
    In contast to outside patent attorneys or agents trying to quickly draft and file applications [often under severe time/cost constraints these days] in many different technologies, many of which they may never have worked on before. [This is affecting quality as well as prior art citations.]

  17. 6

    The Office really needs to repeal Rule 56 and replace it with a rule that one cannot submit English language patents or patent applications, foreign counterpart searchs or office actions, or the like.

  18. 5

    “The applicants duty to disclose should be limited to items that the Office could not or is not likely to find on its own, such as offers for sale and public display of prototypes.”


  19. 4

    Feel the need? There is a duty to disclose, which, if taken literally requires one to submit the entirety of the knowledge of mankind.

    The whole system is based on the assumption that the Office can find prior art. If the Office feels overwhelmed by submissions, the Office can remove the obligation to submit from the applicants and do its own search.

    The applicants duty to disclose should be limited to items that the Office could not or is not likely to find on its own, such as offers for sale and public display of prototypes.

  20. 3

    Point 5 is an interesting supposition and may be an answer to my question. Then again, it may not.

    I am not sure I see either a cause or correlation with point 3.

    For point 2, are you saying that some are submitting references based on an automated process, without actually reviewing the references submitted? Isn’t there a slight ethical problem with such blind submissions?

    Point 1 is included in my Best Practices list (subsumes point 4).

  21. 2

    I suppose:

    1. Those who do pre-filing searches have more docs to cite.

    2. Those who have automated the citation process cite more docs. The tool has been bought and paid for. So, better work it hard.

    3. Those who cite out of an in-house patent facility cite more.

    4. Some have an established habit of citing while others have yet to acquire the habit.

    5. Those that acquire the citing habit are US companies that are trading in a litigious field, like medical devices.

Comments are closed.