What Does DorothyAI’s Scoring Algorithm Have to Say About the Patents Listed in the FDA Orange Book?

Activ8 Novelty

You might be surprised!

The FTC (Federal Trade Commission) recently sent notice letters to 10 pharmaceutical companies, stating that they improperly listed patents in the FDA’s Orange Book. The FTC has determined: “Listing patents in the Orange Book may negatively affect competitive conditions if listings are improper, as defined by law.” The FTC spells out its position in a policy statement published in September, 2023 and an amicus brief filed in the Mylan v. Sanofi filed in November, 2023.

The FTC’s interpretation of the requirements for Orange Book listings are (1) could be infringed by a follow-on drug, and (2) claim either the drug itself (or a formulation for the drug) or an approved method of using the drug. In an effort to determine what Orange Book listed patents meet the statutory requirements for being listed, we compared the independent claims of these patents to the dosage and administration information published in the Orange Book for the related products using DorothyAI’s relevance scoring algorithm. This analysis produced an average relevance score of 75.23 for the independent claims of all Orange Book patents compared to the published dosage and administration. 

At 75.23, the average relevance score for Orange Book listed patents compared to the dosage and administration information published in the Orange Book for the related products is really bad

To put this into context, a score of 0 indicates that the text being compared is identical. Highly relevant patent disclosures typically score at 10 or below. Scores greater than 15 are largely irrelevant. A patent with a relevance score of 75 would not be returned in DorothyAI’s search products, because no patent professional on the planet would conclude that this reference is relevant to the query. In other words, the average relevance score for Orange Book listed patents compared to the dosage and administration information published in the Orange Book for the related products is really bad. 

This is not to say that there are no patents listed that are appropriately relevant. Seventy-two (72) Orange Book listed patents (7% of the total examined) produced relevance scores equal to or better than 15 when compared to the associated dosage and administration disclosure. Three hundred fifty-one (351) Orange Book listed patents (36% of the total examined) produced relevance scores equal to or better than 50. That means that 43% of Orange Book listed patents are probably appropriately listed, assuming that patents with marginally relevant independent claims (scoring 50 or less) may have dependent claims that are more relevant and (1) could be infringed by a follow-on drug, and (2) claim either the drug itself (or a formulation for the drug) or an approved method of using the drug (i.e. the requirements for Orange Book listings). 

On the other hand, 358 Orange Book listed patents (37% of the total examined) produced relevance scores equal to or better than 100, and 170 Orange Book listed patents (17% of the total examined) produced relevance scores below 100 when compared to the associated dosage and administration for related products. It is extremely unlikely that the lowest scoring 17% of Orange Book listed patents meet the requirements as interpreted by the FTC. 

These data suggest that the FTC has some work to do.  

WHAT DOES DOROTHYAI’S SCORING ALGORITHM HAVE TO SAY ABOUT THE PATENTS LISTED IN THE FDA ORANGE BOOK?

Since the dismissive grunts of AI naysayers are interrupting the symphony shocked sighs from our readers, we’ll use Glaxo’s Advair HFA and Flovent HFA as an example. We’ll review the claims of the Orange Book listed patents for these drug products the “old fashioned way” with words and eyes.     

The FTC served notice letters to Glaxo Group Limited for four U.S. patents covering Advair HFA and Flovent HFA. Advair and Flovent contain fluticasone propionate, a glucocorticoid used to treat asthma, inflammatory pruritic dermatoses, and nonallergic rhinitis. Therapeutic formulations of fluticasone propionate were first described in 1989. The claims of USRE40045E1 describe these compositions and methods for using these compositions. USRE40045E1 naturally expired in 2010. 

The independent claims of the Orange Book listed patents covering Advair and Flovent do not cover the composition itself. Instead, these claims cover an “actuation indicator”:

  • U.S. Patent No. 7,500,444 B2:
    1. An actuation indicator comprising a drums sub-assembly comprising a rotatable actuation indicator wheel, a rocking, ratchet pawl for rotating the indicator wheel in a set direction and a rocking mechanism for the pawl driven by a slipping clutch arrangement, wherein the slipping clutch arrangement comprises a slipping clutch spring engaged at one end to a pinion of a rack and pinion assembly and at a second end to the ratchet pawl.
  • U.S. Patent No. 7,832,351 B2:

1. An assembly for an actuation indicator comprising:

(a) an indicator wheel for indicating actuation of a device with which the indicator is to be associated,

(b) a rotatable element of the actuation indicator with which the indicator wheel is engaged, and

(c) an axle spring comprising:

(i) a first axle on which the indicator wheel is rotatably mounted,

(ii) a second axle which extends parallel to the first axle and on which the rotatable element is rotatably mounted, and

(iii) a biasing section connecting the first and second axles to bias them together to bias the indicator member towards the rotatable element.

These claims arguably cover a method for using Advair or Flovent. I do not typically practice in the mechanical arts, but these claims appear to describe fairly rudimentary mechanisms that are definitely not required to administer fluticasone propionate. In my opinion, these claims (1) would most likely NOT be infringed by a follow-on drug, and (2) do NOT claim either the drug itself or an approved method of using the drug. I look forward to hearing Glaxo’s arguments to the contrary.      

How did these patents score during our testing? The average relevance score for Advair HFA and Flovent HFA Orange Book listed patents is 68. We can agree that this is pretty accurate. In fact, this score may be a bit low.   

Notably, Ventolin HFA is the subject of 2 additional Orange Book listings citing the same patents, and producing the same average relevance score. It is unclear why Glaxo was not put on notice for these Orange Book listings. 

To be clear, this is not an indictment of the pharmaceutical companies listing patents in the Orange Book, their legal counsel, or the FDA. Gamesmanship is part of interplay between regulatory agencies and businesses. The Orange Book listings are no exception. We’re here to share the stories the data is telling.

The Orange Book lists the drug approved by the FDA. Pharmaceutical companies list patents in the Orange Book to secure 30 months of exclusivity, blocking lower-cost generic drugs from entering the market and effectively allowing the pharmaceutical company to charge whatever they want for the drug. Generics can challenge the listings at their own expense. Orange Book information was obtained from FDA.gov. Each independent claim of the listed patent was compared to the published Orange Book dosage and administration instructions for the application number associated with the listed patents.