How to read the Orange Book data files

The download is one ZIP holding Products.txt, Patent.txt and Exclusivity.txt, with fields separated by a tilde (~). An application is identified by its type, N for an NDA or A for an ANDA, and its number; each strength is its own product number.

Products.txt lists the ingredient (several are alphabetised and separated by a semicolon), dosage form and route, trade name, applicant, strength, application type and number, product number, therapeutic-equivalence code, approval date, reference listed drug and reference standard flags, type (RX, OTC or DISCN) and the applicant's full name. Products approved before 1 January 1982 carry the words "Approved prior to Jan 1, 1982" in place of a date; the app keeps the words and gives no date for them.

Each row gives the application and product, the patent number, the patent expire date, the drug substance and drug product flags, the patent use code, the delist request flag and the submission date. The expire date is the date as the holder submitted it, including extensions; the submission date is the date FDA received the patent information.

The substance and product flags apply only to patents submitted on Form 3542 and listed after 18 August 2003, and each is Y or empty. An empty flag on an older listing does not mean the patent claims neither the substance nor the product.

A use code marks a patent that claims a method of use for an approved indication; the holder must describe the specific approved method of use it claims (21 CFR 314.53(c)(2)(ii)(P)). One patent can appear on several rows, one per product number and one per use code.

Holders list drug substance, drug product and method-of-use patents, and must not submit process, packaging, metabolite or intermediate patents (21 CFR 314.53(b)(1)), so those are not meant to appear in the files.

FDA's description of the patent number field says it includes pediatric exclusivity granted by the agency. In the file such a row is the patent number followed by *PED, with its own later date and empty flag columns.

Pediatric exclusivity is a six-month bar on approval counted after the patent's date (21 USC 355a(b)(1)(B)), not a longer patent term. The app shows the *PED row's date as FDA lists it, reads the substance and product flags from its base rows where they agree, and never adds six months itself.

The delist request flag means the holder asked for the patent to be removed, but it stays listed under the 180-day rules. FDA removes such a patent only when no first applicant is eligible for 180-day exclusivity based on it, or after that period has expired or been extinguished (21 CFR 314.53(f)(2)). The app keeps these rows and flags them.

Each row gives the application and product, an exclusivity code and the date the exclusivity expires. The meaning of each code is FDA's; the app shows it from the code list it keeps, and an unknown code is shown as unknown. The Preface notes that a waived exclusivity is coded W.