George Mason University Antonin Scalia Law School

Pharmaceutical “Nominal Patent Life” Versus “Effective Patent Life,” Revisited

By Emily Michiko Morris and Joshua Kresh

Overlaid images of pills, a gloved hand of someone expecting a pill, and an eyedropperExecutive summary: Many critics of pharmaceutical companies argue that they abuse the patent system through “evergreening” or “thickets” to increase the amount of time they can avoid generic competition and keep drug prices high. Those critics have not looked at the real-world effects of pharmaceutical patents on generic entry, however. Our review of actual time to generic entry for more than one hundred of 2012’s top-selling drugs shows that:

    • The average effective patent life, as opposed to nominal patent life, of our dataset is 13.35 years, consistent previous studies on effective patent life;
    • Patents and exclusivities added to the Orange Book after a drug’s market entry does little to extend effective patent life; and
    • The number of patents protecting a brand-name drug has no significant correlation with effective patent life.

Thus, our study suggests that “evergreening” does not stop generic entry and that “thickets”—if they even exist—appear to be rather easy to circumvent.

The topic on everyone’s minds lately is drug prices and the fact that most Americans believe that drug prices in the United States are too high. Drug prices, like other health care costs, are a multifactorial and incredibly complex subject. Most of the current discussion on drug prices focuses on the role of patent protections, however, to the exclusion of almost everything else. In particular, a major criticism of the pharmaceutical industry is that it is abusing the patent system by filing for serial patents to prolong its ability to charge supracompetitive prices for the drugs that it has developed.

To prove the existence of such “evergreening” through patents, a number of studies focus on nominal patent life, based on the expected expiration date of the last patent on a given set of drugs. The later the expiration date, according to evergreening theory, the longer a brand-name drug can fend off entry by price-lowering generic versions. The most well-known—and certainly the most thorough—study applying this approach is Prof. Robin Feldman’s “Evergreen Drug Patent Database” (often informally referred to as the “Hastings Database,” after UC Law San Francisco’s former name).

The Hastings Database contains an exhaustive list of not only all patents but also any regulatory exclusivities granted by the FDA, both of which can stall generic drug approval and thus market entry as well. The Database then identifies “evergreening” by looking at how many additional patents or exclusivities are added to the “Orange Book,” the FDA’s list of patents and exclusivities that pharmaceutical companies assert cover their brand-name small-molecule drugs (SMDs).[1] Specifically, the Hastings Database counts how many patents and exclusivities are added after what the database labels as the “protection cliff” for each drug, as defined by all patents and exclusivities added to the Orange Book by two months after FDA approval.[2] According to the Hastings Database’s calculations, companies extend the patent lives of their drugs for several years by adding such later filed patents and exclusivities.

This calculation presents merely the nominal patent life of a given drug, however, not the period of time during which patents actually protect a drug from generic market entry. The latter, or effective patent life, is a more accurate and more meaningful measure of how long brand-name drug companies can fend off generic market entry. Unlike nominal patent life, effective patent life (EPL) does not focus on patent terms. Instead, EPL focuses on the time between a brand-name drug’s approval and first market entry and generic market entry because this is the only time in which the brand-name company might be able to charge supracompetitive prices. The differences between nominal patent life and effective patent life can be quite large, as generics often enter the market regardless of whether the brand-name still has patent term remaining.[3] This is a point that many earlier studies have shown.[4]

To reaffirm this point, we did our own study of the nominal patent lives listed in the Hastings Database. For our sample set, we looked at the top-selling small-molecule drug products from 2012, based on the idea that flagship brand-name products are most likely to draw generic market entry and that drugs from 2012 would now have had twelve years in which generics could do so. To select the drug products for our sample, we used the list of the top 200 drugs by total U.S. retail sales in 2012 assembled by the Njardarson Group at the University of Arizona.[5] After we eliminated any biologics, as well as any SMDs not included in the Hastings Database, we had a sample size of 131 drug products.

We then added data from the Hastings Database. These data included each drug product’s FDA approval date, the expiration dates for both the earliest and latest patent or regulatory exclusivity listed in the Orange Book for each drug product, and each product’s “protection cliff” dates. We also included any further time past those protection cliff dates that the Hastings Database identifies added by patents or exclusivities beyond those that comprise each protection cliff. This latter set of data, which the Hastings Database labels as “Additional Prot(ection) Time,” is important because it is how Hastings calculates alleged “evergreening.”

To these data from the Hastings Database we then added data from other resources as well: both the date on which each Reference List Drug (RLD) began marketing (i.e., the date on which the relevant brand name entered the market), and similarly the date on which the first generic for each RLD entered the market. We added these market entry dates from the earliest listed dates included in the National Drug Code Directory’s Structured Product Labeling Resources (SPL) database[6] for the earliest approved New Drug Application (NDA) listed in the Hastings Database. (A large number of drug products have multiple NDAs and multiple market entry dates, so we used the earliest RLD market entry dates for the earliest approved NDAs to err on the side of the longest EPLs for each product.) Based on these dates, we calculated the EPL for each drug product based on the time between the product’s first marketing date and the date on which the first generic for that product entered the market.

Because the Hastings Database defines evergreening as protection beyond its “protection cliff” rather than as nominal patent life, we computed two further datapoints. First, to compare directly with the Hastings Database’s “Additional Prot(ection) Time,” we also calculated EPL based not on RLD market entry dates but on Hastings’ protection cliff dates—that is, we calculated the effective patent life for each drug product beyond its protection cliff. This allowed us to compare what is in effect Hastings’ nominal “Additional Prot(ection) Time” with what is in effect our sample’s effective “Additional Prot(ection) Time.” Second, to make the Hastings Database more comparable with nominal patent life (NPL) determinations in other studies, we also derived the NPL for each drug product based on the time between its RLD market entry date and the latest expiration date of any patent or exclusivity listed in the Hastings Database for the product.

Our analysis of our sample set is ongoing, but some of the initial results are significant. Not surprisingly, the average EPL from our sample—including the 14 drug products for which the FDA currently lists no generic versions—is several years shorter than the average NPL we computed from the Hastings Database. The average NPL from Hastings is 19.14 years (median = 19.20), but the average EPL from our sample is 13.35 years (median = 14.01). Our sample’s average EPL is thus consistent with EPLs from other studies.

More interesting, however, is that our sample’s average effective “Additional Prot(ection) Time”—1.61 years (median = 1.19)—is also much shorter than Hastings’ nominal “Additional Prot(ection) Time”—13.34 years (median = 13.52). In other words, the effective patent life of our sample, on average, extends only 1.61 years past Hastings’ “protection cliff.” This means that most of the mean EPL from our sample stems from the patents and exclusivities that comprise Hastings’ protection cliff (those listed in the Orange Book up to two months after FDA approval). This in turn shows that if, as is frequently claimed, patents and exclusivities are later added for brand-name drug products simply to avoid their protection cliffs, that particular tactic is ineffective.[7]

That being said, many of the drug products in our sample may have had shortened EPLs because generics were able to enter the market early through Paragraph IV certifications contesting either the infringement or validity of the latest expiring patents for those drugs. We therefore looked at the approval letters for as many of the earliest entering generics as we could find on the Drugs@FDA: FDA-Approved Drugs online database. The FDA’s approval letters typically include whether the approved generic has filed a Paragraph IV challenge and which patents it were challenging. We were able to pull up generic approval letters for 87 of the drug products in our sample. Of those products, 16 either faced no Paragraph IV challenges at all or at least none challenging the latest expiring patent. For another 26 of the 87 drug products, their patent owners did not sue the first-to-file generic even though the generic filed a Paragraph IV challenge to the latest expiring patent. This does not mean that the first-to-file generic did not itself then file a declaratory judgment action against the latest expiring or other patents, but it does mean that the patent owner did not think it worthwhile to sue the Paragraph IV generic early enough to obtain a 30-month stay on that generic’s FDA approval. Several of the 87 products, however, had multiple first-to-enter generics entering the market on the same day, but the FDA database did not display the approval letters for all those generics. We may therefore be underestimating the number of products that faced Paragraph IV challenges.

It is also possible that the time needed to resolve Paragraph IV challenges by itself may have delayed generic entry in many cases. Similarly, it is possible that the mere existence of later-expiring patents deterred potential generics from even trying to enter the market early. We therefore used Hastings’ raw data to derive the number of patents protecting each drug product in our sample. We counted all individual patents listed, treating any pediatric extension, patent term restoration, or other patent term extension as a separate patent if it had the potential to extend nominal patent life. We then looked for any correlation between the number of patents per drug product and the effective patent life for each product but found no statistically significant difference from a null hypothesis of zero correlation. This again suggests that simply adding more patents, regardless of whether they are listed in the Orange Book later or earlier, is not an effective tactic for delaying generic market entry.

Perhaps most significantly, our findings suggest once again that looking at only patents and patent terms reveals little to nothing about how long brand-name drug products can stave off generic entry. Nominal patent life, for example, tells us little about the actual effect patents have because nominal patent life fails to consider the scope of each patent. Many patents, especially later-filed patents, on new indications for which a drug patent can be used or new ways of manufacturing a product, can either be carved out of a generic’s FDA application or designed around. Even new dosage patents may not stop generic entry if physicians can simply split or multiply the dosage of a generic to achieve the newly patented dosage. Much the same can be said of new formulation patents as well. And even if other types of patents can be avoided only through Paragraph IV challenges, these challenges may have little effect in extending effective patent life, as suggested by our data.

[1] Small-molecule drugs are small and simple substances that can be synthesized though chemical reactions, unlike “biologics,” which are a relatively new class of therapeutics that are much larger and more complex molecules that are synthesizable only through biological processes.

[2] The Hastings Database also calculates for each drug the length of time between the expiration of its first patent or regulatory exclusivity and the expiration of its last.

[3] See, e.g., C. Scott Hemphill & Bhaven N. Sampat, When Do Generics Challenge Drug Patents?, 8 J. Empirical L. Stud. 613, 643 (2011) (noting that effective patent lives are shorter than nominal patent lives).

[4] See, e.g., Henry G. Grabowski et al., Continuing Trends in U.S. brand-Name and Generic Drug Competition, 24 J. Med. Econ. 908, 916 (2021) (calculating to EPL – or “market exclusivity period” (MEP) – as only 13.0 to 14.1 years for new chemical entities); C. Scott Hemphill & Bhaven M. Sampat, 31 J. Health Econ. 327, 330 (2012) (finding EPL of 12.15 years versus NPL of 15.89 years for new chemical entities).

[5] chrome-extension://efaidnbmnnnibpcajpcglclefindmkaj/https://sites.arizona.edu/njardarson-lab/files/2023/11/Top-200-Pharmaceutical-Products-by-US-Retail-Sales-in-2011_small_0.pdf.

[6] https://www.fda.gov/industry/structured-product-labeling-resources/nsde

[7] The difference between average EPL and NPL for the products in our sample is statistically significant, based on a paired two-tail t-test with p value <<0.01. The same is true of the difference between effective and nominal “Additional Prot(ection) Time” for our sample.