Rounding Errors in Patent Law

by Dennis Crouch

Astrazeneca AB v. Mylan Pharmaceuticals Inc. (Fed. Cir. 2021)

This case helps show that patent law has arrived at Nerdsville.  Actually, I’ve been here for a while and so welcome to all you newcomers.  In my view, the case also suggests that our claim construction doctrine has gone too far attempting to discern meaning where none exists.  Rather, I would suggest that the issues here are really about infringement rather than claim construction.

The claims require “0.001%” of a formulation be the inactive ingredient PVP K25.  The allegedly infringing formulation is slightly different — lets say 0.0008% — but it is within standard rounding error if 0.001% is seen as having one significant digit (0.0005 to 0.0014).  Mylan argued that 0.001% should be given a narrower range — “the precise number, with only minor variation.”  Mylan’s outcome would result in a range of something like 0.00095 to 0.00104.

The district court sided with the patentee — holding that standard scientific notion indicates that a physical measure such as this includes the rounding error according to the number of significant digits shown. If they had wanted a more precise number, then scientific notation permits the addition of more significant digits — such as 0.0010%.  On appeal though the majority rejected that approach based upon statements from the specification comparing 0.001% against 0.0005% (a value that should be within the rounding error).

Thus, taken as a whole, the intrinsic record supports a narrower construction of 0.001% to reflect that term’s application to the PVP concentration in particular, and the testing evidence in the written description and prosecution history showing that very minor differences in the concentration of PVP—down to the ten-thousandth of a percentage (fourth decimal place)—impact stability.

Slip Op.  The court noted that the patentee could have used the term “about” to extend the range.  Judge Stoll wrote the majority opinion that was joined by Judge Hughes.

Judge Taranto in dissent first noted the uncontested conclusion that the ordinary meaning follows the broader rounding range:

The significant-figure convention for real-world-measurement situations supplies an answer by giving a well-defined interval as the ordinary meaning, in the art, of a statement of a single number like the statement at issue here. Under this ordinary-meaning approach, the significant-figure interval meant by “0.001%” is 0.0005% to 0.0014%, based on rules of rounding and the single significant figure at the third decimal place.

Dissent.

Judge Taranto then argues that the proposed “minor variations” limit creates confusion because it lacks precision.  Of course, we’re talking about rounding and measurement errors here and so they already lack precision.

This video was suggested by a friend as relevant:

38 thoughts on “Rounding Errors in Patent Law

  1. 10

    In Ortho-McNeil Pharma, Inc. v. Caraco Pharma Labs (Fed. Cir.2007) patent refers to pharmaceutical composition comprising a weight ratio of two drugs known as analgesics in which said ratio is “about 1:5”. The Court concludes that the term “about” (about) has the same meaning as “approximately” (approximately). The term should be interpreted in terms of the technological area. In this case, the Court concludes that the statistical variability of the measurements could assume values ​​in the range from 1:3.6 to 1:7.1. In Cohesive Tech Inc v. Waters Corp (Fed. Cir 2008). The patent deals with the process of separating a compound from a liquid using particles with an average diameter greater than about 30 micrometers. Again tests in the area point to a measure variance of around 15%, which means a range between 25 and 34 micrometers. In BJ Serv Co. v. Halliburton Energy Services (Fed. Cir 2003) the claim refers to a concentration value of approximately 0.06%. The controversy over the term “approximately” can be resolved by proving that this value was the normal range of experimental error for measuring this parameter.

  2. 9

    I want to pick up on Greg’s thoughts expressed at #2, on the issue of how to construe a point value in a claim, and then how to decide whether an accused infringement is “equivalent”. I do it in the context of an alloy composition but I think it applies to any commercially manufactured chemical product.

    In short, the maker cannot make the product with the point value of the ingredient exactly on the nail. There will always be a small statistical variation about that target value of, say, nitrogen in steel or nucleating element in a dispersion-hardened aluminium alloy composition (unless you want to count the atoms one by one).

    So, the court when construing the claim cannot duck out of the job of giving the point value a range, in order to decide what is a literal infringement and what is an infringement by equivalent.

    The UK Supreme Court, when construing a claim, asks itself “What was the inventor using the language of the claim to mean”. Courts in the USA ought to do the same.

    1. 9.1

      I think that you miss the ‘point’ here that — without any need for judicial input — point values already carry that range based on the common understanding of “significant digits.”

      To wit, the value of “0.01” is NOT precisely the same as the value of “0.0100”. Each of these distinctly already have an understood — and different — “inherent range.”

      Anyone with any sort of technical background should see this a mile away.

      1. 9.1.1

        Well yes, I do know about places of decimals, and significant figures. But isn’t the point here that this was not the end of the matter? Instead, the court majority borrowed from the specification to displace that construction and arrive at a different construction of the claim.

        And as for the court’s encouragement to claim drafters to fill up their independent claim with the word”about”, I’m flabbergasted. How much difference would it make? I mean, English courts would say that when reading a claim, out of fairness to the inventor, it is often right to take it that there is an implicit “about” prefacing every numerical point and range.

        1. 9.1.1.1

          Color me shocked that a court would intrude on a scientific (factual) matter and muck things up.

          Again, it is well understood what “implicit about” means. Let me refer you again to:

          To wit, the value of “0.01” is NOT precisely the same as the value of “0.0100”. Each of these distinctly already have an understood — and different — “inherent range.” (or “implicit about”)

    2. 9.2

      As to your prescriptive ‘here, wear my glasses’ comment of, “The UK Supreme Court, when construing a claim, asks itself “What was the inventor using the language of the claim to mean”. Courts in the USA ought to do the same.”…

      May I offer you some burnt biscuits?

      1. 9.2.1

        Indeed, anon. It was with the Chef America case in mind that I wrote that courts in the USA should ask themselves what was the drafter of the claim using the language of the claim to mean. Very obviously NOT a heap of charcoal but, rather a delicious fresh-baked loaf of bread. Any technical reader, any person skilled in the art, would grasp that immediately.

        But then, how much easier is it for a judge to don the mantle of a skilled philologist/grammarian, and divine the meaning for such a person of the claim in view. Much easier. By far. Surely. So much more scope for getting it wrong, if you make the mistake of trying to construe the claim through the eyes of the intended reader. So much more work, so much more hard concentration. So much more to read and digest. All that work, and risk. The horror of it all. Judicial economy far more attractive.

  3. 8

    “This case helps show that patent law has arrived at Nerdsville.”

    I’ll wager that the rest of legal profession has for at least 100 years considered patent law to be Nerdsville (or whatever the equivalent term of “nerd” was 100 years ago).

  4. 7

    Rather than using a precise number and hoping a court would imply a range of values in their favor, it would have been (probably) a better move for the patentee to simply put a range of values in the claim. Then there is no question as to what the claim covers or does not cover.

    1. 7.1

      Agreed. That appears to be a point along the same lines that #5 & #2 below are each making, and it sounds right to me.

    1. 6.1

      Of course we would brosef. But corporate dems are way too worried about stuffing the ballot box, and corporate repubs are way too worried about not having the ballot box stuffed against them. There is no will and desire among the overclass to legitimize by law the creation a large voteless underclass (the obvious solution which is being touted by many would-be legalized dreamers and other illegally present migrants), and so there is no solution. Politically.

      Setting that aside, we’re leaving money on the table about a million other ways as well.

      1. 6.1.1

        [W]e’re leaving money on the table about a million other ways as well.

        True, and good point. I am broadly in favor of picking up those other piles of money (zoning reform, FDA reform, etc) as well. Failure to pick up pile of money A, however, is not a good reason to fail to pick up pile B.

        1. 6.1.1.1

          I’m not sure how zoning reform is picking up piles of money, as it is rather generally speaking TAKING piles of money from individuals which is why it isn’t done more (people stop themselves from getting robbed).

          But in any event, the overall selling out of the country for a dollar by the boomers (ahem Greg), is nothing new. Boomers get theirs, may as well turn the whole place into the 3rd world for a dollar, you know, except for the boomer’s estates, which will remain single fam zoned, before the boomers die off so everyone else gets to pick up the pieces. Boomers-not even once.

          Anyhow, you bros excited to see Roe going? I can’t believe I correctly predicted this before Kman got to be a justice. Truthfully I didn’t see it happening quite so quickly. Trumpadump the carny making a move from beyond the presidential grave. But, either way, I’ll take the W.

          link to youtube.com

          1. 6.1.1.1.2

            Incidentally, it is incumbent homeowners (like most boomers), who lose out from zoning reform. The idea that zoning reform is a sort of boomer plot betrays rather a serious misunderstanding of the policy stakes at issue.

            1. 6.1.1.1.2.1

              “Incidentally, it is incumbent homeowners (like most boomers), who lose out from zoning reform. The idea that zoning reform is a sort of boomer plot betrays rather a serious misunderstanding of the policy stakes at issue.”

              Uh yeah, that’s what I said. The boomer “plot” is to signal their virtue by inviting the third world into the country to increase pop to keep SS afloat and “growth” continuing without having babies, and as a side effect driving up home prices (which they own most of), and then signal their virtue still more by getting zoning reform done for people other than themselves, and then NIMBY stopping the zoning reform in their own neighborhoods (preventing them from losing money). All this leaves the boomers HIGH in virtue, and HIGH in home value (also high in SS/retirement $$$). They die off, everyone else has to pick up the pieces from their signaling of virtue and having championed zoning reform for everyone outside of their little neighborhoods. Not really an explicit “plot” written down by the head boomer or anything, it’s just the net effect “plot” of the policies that they are enacting. (I really want virtue, but I also really want home value for me, someone else can pay for the virtue!) And alot of boomers are aware of it too, the news isn’t completely silent. My uncle for example knows good and well how his generation is robbing the other generations via various means.

            1. 6.1.1.1.3.1

              I’m not against them for their age anon, I’m against them for their actions as a collective age-group robbing others so they can live the high life.

              1. 6.1.1.1.3.1.1

                So you favor zoning reform then? Single family zoning (and NIMBYism in general) is one of the principal channels by which they pick society’s pocket (one one easily remediable within their lifetimes).

      2. 6.1.2

        “and so there is no solution. Politically.”

        It looks like you’re trying to point at “both sides” while carefully walking around how exactly one side is imposing the “voteless” condition.

        1. 6.1.2.1

          Ben,

          I head an analogy recently that fits here: a bowling alley.

          One gutter is that “voteless” state.
          The other gutter is the “votes for non-(proper) citizens” state.

          As 6 points out – this IS a ‘both-parties’ issue.

        2. 6.1.2.2

          “It looks like you’re trying to point at “both sides” while carefully walking around how exactly one side is imposing the “voteless” condition.”

          Not sure what you’re talking about, neither “side” is currently imposing the voteless condition, that is the law from yesteryesteryesteryear not being properly enforced that is creating the voteless condition on around 10-30 milly people today in the US. It is the lack of enforcement that is doing it today. But are you referring to Dems trying to stuff the ballot box? Or what?

          1. 6.1.2.2.1

            I guess I misunderstood you. It seemed like you were saying there was a solution for the dreamers of allowing them to stay legally without any path to citizenship, and that neither side was interested in such a solution. If that wasn’t your point, sorry.

            1. 6.1.2.2.1.1

              You understood me correctly. There is such a hypothetical path.

              “and that neither side was interested in such a solution”

              They’re not. Remember, these are people “on both sides” who are the ev il ones, those of the “whiteness as a social construct”, they who still believe in the ev il freedom of the Angleishman from history yore, where the right to a vote is synonymous with not being a serf/slave.

              What does that have to do with one side or the other supposedly imposing the voteless condition?

      3. 6.1.3

        “we’re leaving money on the table about a million other ways as well.”

        Also, I think this is wrong in an important sense.

        Yes, we leave money on the table in various ways. And China is surely leaving similar money on the table in similar ways.

        But China doesn’t have the opportunity to drain the brains of the world that we do. When we waste that opportunity, we are wasting a competitive advantage that we have. It would be like if China wasted it’s authoritarian government advantage by having massive governmental turnover every 2-4 years. We’re accepting a handicap on one of our best events of the competition.

        This is not just leaving money on the tale. It’s leaving our edge on the table.

        1. 6.1.3.1

          “But China doesn’t have the opportunity to drain the brains of the world that we do.”

          He said as china imports/drains the brains of a lot of large corps irl.

          “This is not just leaving money on the tale. It’s leaving our edge on the table.”

          ^Reddit tier comment.

  5. 5

    This is ridiculous either way. If they wanted the small range then they could have claimed such. If the court goes on and on about allowing small variations then you have to compare all the prior art that would be within those small variations against the claim as well.

  6. 4

    Judge Stoll and Judge Hughes need to be required to sit through Chem101 again. Most first year students are able to get this one right.

    1. 4.1

      Given his background, I seriously doubt Judge Hughes ever took a chemistry class at any level.

  7. 3

    I am struck by the order of magnitude issue (unstated) here.

    When the ‘math’ is so loosely tossed around, the reality of the numbers becomes a bit a of a blur.

    We are talking about an INERT element at 9what exactly) percentage…?

    The phrase “get real” seems apt.

    1. 3.1

      Inactive ingredient (any ingredient except the active ingredient) doesn’t mean inert (as in completely chemically etc. inactive) here bro. The active ingredient is the one that is biologically active in specific. They should just call it the biologically active ingredient. There are lots of inactive ingredients in various chemical formulas and drugs etc. that are not what you are thinking of as inert. Preservatives, delivery mechanisms for the active ingredient, etc. etc.

      link to verywellhealth.com

  8. 2

    I think that the majority got it right here. The reason one claims a precise number like this (instead of a range) is because there would be a problem (enablement, prior art, etc) with the broader claim. In such circumstances, it is right and just to hold the literal infringement analysis to the claimed value.

    Perhaps there is a doctrine of equivalents issue to consider on remand. As a matter of literal infringement, however, the district court’s construction was wrong. The ANDA applicant is entitled to a remand.

  9. 1

    “I would suggest that the issues here are really about infringement rather than claim construction.”

    It certainly worked to Mylan’s advantage for it to be the latter. That way it got de novo review on appeal, whereas for the former it would’ve been substantial evidence review of a jury verdict.

    The opinion also states that Mylan stipulated to infringement after Markman, so it got a bench trial on its invalidity claims instead of putting them to a jury.

    Speaking of “Nerdsville” when I view this post in Chrome on a desktop, some of the text wraps to the left side of the “Rounding Error” graphic. The problem isn’t present on a phone browser though.

Comments are closed.