Estimating The Cost Of Delayed Generic Drug Entry To Medicaid

Pharmaceutical spending in the US continues to rise at rates that outpace those in other industrialized nations, reaching $453 billion in 2017 (not including rebates) and accounting for 17 percent of health care spending.¹ When a brand-name product enters the US market, patents and other statutory market exclusivities allow its manufacturer to charge high prices, unchecked by competition.² This exclusivity period lasts a median of about 12.5 years,³ with several factors contributing to its duration.⁴

The key patent on a drug, which usually protects from others using the same active ingredient (that is, the chemical compound or therapeutic moiety itself), must generally be obtained near the time of drug discovery. As a result, some years of patent protection are expended during clinical investigational testing and regulatory review before the drug is approved by the Food and Drug Administration (FDA). To offset a portion of this time, brand-name manufacturers are granted up to five years of “patent term restoration” on a single key patent of their choosing, extending its duration by up to five years—to a maximum of fourteen years from the date of FDA approval.⁵ Previous research has shown that patent term restoration is sought for over 90 percent of approved drugs⁶ and that 93 percent of these restored patents are for the active ingredient.⁷ Patent exclusivity can be further extended by six months if manufacturers conduct trials of their drugs on children.⁸ The combined effect of these two policies is to provide market protections for new pharmaceutical products for up to 14.5 years (a statutory limit that has been reached by 73 percent of drugs with patent term restoration extensions).⁷

Brand-name manufacturers have strong financial incentives to prolong market exclusivity and delay generic entry as long as possible, even beyond the initial market protections discussed above. One approach is to obtain additional patents on other aspects of the product, such as for methods of use, alternative chemical formulations, or the delivery device.⁹ The delay then arises as generic manufacturers challenge these patents in court or an administrative forum as being improperly granted or not applicable to their generic versions. Other strategies include withholding drug samples from generic manufacturers that they need to complete FDA-required bioequivalence testing or filing a citizen petition with the FDA, arguing that generic versions should not be approved because of safety concerns.¹⁰

Because generic competition reduces drug prices, delays in generic drug entry can result in excess spending by payers and patients. To understand the financial impact of delayed availability, we identified a cohort of drugs predicted to lose market exclusivity and determined when a generic version actually entered the market. When generic entry was delayed, we estimated the excess costs incurred by Medicaid.

Study Data And Methods

Cohort Selection

The US Patent and Trademark Office (USPTO) keeps a list of drugs granted patent term restoration, including information on which drugs were granted restoration and which patents were restored. Using this database, we identified a cohort of drugs with key patents that expired in the period 2010–15. To ensure that other statutory exclusivities did not outlast patent exclusivity, we cross-referenced this list with information on the FDA website to ensure that exclusivity under the Orphan Drug Act of 1983 (which lasts seven years for drugs to treat rare diseases) or any other statutory exclusivity did not outlast the patent expiration date. Finally, we added six months to the exclusivity period if the manufacturer had responded to a written request to conduct pediatric clinical trials.

We excluded biologic medications because the first interchangeable “generic” (biosimilar) version of a biologic drug was approved by the FDA in 2015. We also excluded drugs that were converted to over-the-counter medications (for example, orlistat) and drugs primarily used in inpatient settings that would not be observable in the Medicaid outpatient data set (such as intravenous antibiotics).

Medicaid Data And Identifying Delays In Generic Entry

This study used data for 2010–16 from the Medicaid State Drug Utilization files published by the Centers for Medicare and Medicaid Services. Because our cohort comprised drugs expected to lose market exclusivity between 2010 and 2015, using data up until 2016 allowed at least one year for generic entry to occur from the date of expected loss of market exclusivity. The files are updated quarterly and include medication use information for Medicaid plans compiled at the state and National Drug Code (NDC) levels.¹¹ Relevant data elements include the total number of units dispensed, Medicaid and non-Medicaid reimbursement amounts, NDC numbers, and drug names. However, the drug name field is truncated, so information on the active ingredient and administration form is often missing. Using the NDC number, we merged these files with the Red Book database, a drug information aggregator for over 200,000 pharmaceutical products. This provided additional drug-level information on data elements related to full drug names (including a drug’s active ingredient, strength, and route of administration), manufacturer name, and whether the dispensed medication was a brand-name or generic drug.

For each drug we estimated the difference between the expected and actual quarter-year of generic drug entry. The actual quarter of generic entry was defined by the first appearance of a generic equivalent of the brand-name product in the Medicaid data. A delay in generic entry was defined as either the time between the expected and actual generic entry for products that eventually received a generic version or, for cases in which there was no generic entry, between the expected generic entry and the last quarter of 2016 (the end of the study data).

Medicaid Spending Associated With Delays

Medicaid spending before rebates on brand-name products that experienced a delay in generic entry was estimated during the delayed quarters. All drug prices were inflated to 2016 dollars using the Consumer Price Index. For data from fee-for-service plans, we computed the total spending for the delayed drug. Because managed care plans within Medicaid incompletely capture reimbursement data,¹¹ we used the fee-for-service data for each drug and divided the reported total reimbursement by the units dispensed to estimate the per unit brand-name drug reimbursement price (for example, the price per tablet dispensed). We then multiplied this estimate by the total units dispensed for the managed care plans (for example, the number of tablets dispensed). Total Medicaid spending was calculated by summing the estimates from both the fee-for-service and the managed care Medicaid plans.

For the delayed quarters, we calculated the excess spending by determining the difference between the estimated Medicaid spending had generic entry occurred during the expected quarter and the actual Medicaid spending, accounting for three parameters (see online appendix exhibit A1 for additional details on the first two parameters).¹² The first parameter was brand-name market-share erosion, which signifies the extent to which brand-name products are dispensed as a percentage of the overall brand-name and generic drug market. Based on previous research findings,³ we assumed that a brand-name drug retained 25 percent of its original market share in the first quarter of generic entry, 20 percent in the second quarter, and 15 percent in the third and fourth quarters. Because data on long-term brand-name market share are not available, we assumed a conservative 10 percent brand-name market share in periods beyond one year after generic entry.

The second parameter was the ratio of generic to brand-name prices, which was estimated using data from a prior report that calculated the pre-rebate ratios for a cohort of generic products that entered the market in the period 2011–13.¹³ We assumed that a generic drug’s price was 83.5 percent of the brand-name drug’s price in the first quarter after generic entry, decreasing to 34 percent and 26 percent by the end of the first and second years, respectively (see appendix exhibit A1).¹²

The third parameter was drug rebates. Because product-level rebate data are confidential, we used the latest available aggregated reimbursement data to assume 60 percent and 15 percent Medicaid rebates for brand-name and generic drugs, respectively.¹⁴

To assess the robustness of our study findings, we also conducted two sensitivity analyses. First, because brand-name rebate levels are the most important determinants of excess spending associated with delays in generic entry, we varied the percentage of brand-name rebate from the base-case scenario of 60 percent to between 0 percent and 90 percent (in increments of 10 percent). Second, we assumed a 1:1 ratio of generic to brand-name prices for the first six months of generic drug entry, to account for the 180-day exclusivity period (a period of market exclusivity granted to the first successful generic applicant that files an application with the FDA that challenges a brand-name drug patent).¹⁵

Reasons For Delay In Generic Entry

For the fifteen drugs that accounted for the majority of the total Medicaid spending, we attempted to ascertain the reasons for delays in generic entry. First, we queried the commercial pharmaceutical database AdisInsight, which is widely used by pharmaceutical researchers to identify US patent and exclusivity information for a given drug. We also examined the FDA Paragraph IV Patent Certification, Westlaw Edge, and UnifiedPatents databases for additional information on patent litigation and other administrative patent challenges.

Limitations

Our findings should be viewed in light of several limitations. First, the findings were based on most recently published aggregated rebate amounts, as product-level rebate data are confidential.

Second, the findings are generalizable only to Medicaid plans (since the data are sourced from Medicaid files) and to medications commonly used in an outpatient setting. As our study did not consider inpatient drugs, biologics, or drugs for which generic entry was delayed before 2010, the actual excess spending is assuredly higher. These estimates also did not account for additional spending related to the lack of generics for reformulated or combination products, which has been estimated to be in the hundreds of millions of dollars annually.¹⁶

Third, we used parameters from the literature to inform our parameters for brand-name market share and ratios of generic to brand-name prices, and all scenarios assumed a marked reduction in prices upon generic entry. However, given the current controversies surrounding cases of high generic prices—a phenomenon often attributed to low competition levels^17,18 and drug shortages¹⁹—this assumption might not hold for some products.

Fourth, some of the patents that caused delays in generic entry may have been granted for clinically relevant innovations.

Fifth, the study did not consider the cost savings in instances when generic drug entry occurred before the expected entry date.

Study Results

In the period 2010–15, sixty-nine brand-name products were expected to lose market exclusivity (see appendix exhibit A2 for the entire list of products included in the study, along with the expected quarter of loss of market exclusivity).¹² The most common therapeutic categories represented in the study were drugs that affected the central nervous system (thirteen drugs, or 19 percent of the sixty-nine products), anti-infectives (nine drugs, or 13 percent) and drugs that affected the cardiovascular system (eight drugs, or 12 percent). Other common categories were drugs that affected the eyes, ears, nose, or throat; those that affected the autonomic nervous system; and antineoplastics. There were seven drugs (or 10 percent) in each of these three categories.

Eleven (16 percent) of the sixty-nine products had not experienced the entry of a generic product by the end of 2016. Thirty-eight products (55 percent) experienced generic entry either before or within one quarter of the expected date. Twenty-six products (38 percent) experienced generic entry as expected, while twelve products (17 percent) experienced generic entry earlier than expected (exhibit 1). For twenty products (29 percent), generic entry was delayed by more than one quarter-year. For these products, time to generic entry varied substantially, but entry was more likely to occur more than one year after the expected date.

SOURCE Authors’ analysis of data for 2010–16 from the Medicaid State Drug Utilization files. NOTES For details on how the expected and actual generic entry quarters were determined, see the text. “Early entry” means that entry was detected more than one quarter-year prior to the expected date, which was the case for twelve (17 percent) of the drugs. “As-expected entry” means that entry occurred within one quarter of the expected date, which was the case for twenty-six (38 percent) of the drugs. “Delayed entry” means that entry occurred more than one quarter after the expected date, which was the case for twenty (29 percent) of the drugs. The remaining eleven (16 percent) of the drugs did not enter the market in the study period.

Costs Associated With Delays In Generic Entry

There were thirty-one brand-name products for which generic entry either was delayed by more than one calendar quarter or did not occur (exhibit 2). The study identified darunavir (Prezista) as the product associated with the highest share of excess Medicaid spending due to delays in generic entry. Although a generic version for darunavir was expected by the first quarter of 2015, none was detected during the study period (which ended in the fourth quarter of 2016). After accounting for rebates and using market average predictions of brand-name market share and ratios of generic to brand-name prices, we estimated that excess spending during the period of delayed market entry was $137 million. Generic versions of paliperidone (Invega), atorvastatin (Lipitor), and olopatadine (Pataday) were delayed for 3.00, 1.75, and 4.50 years, respectively, and were associated with estimated excess Medicaid spending of $126 million, $62 million, and $100 million. The total estimated excess Medicaid spending related to delays in generic entry for the thirty-one drugs with delayed or no generic entry was $761 million over six years, with the top five drugs alone contributing nearly two-thirds to this estimate.

In the first sensitivity analysis, after we accounted for market average estimates for ratios of generic to brand-name prices and brand-name market share, excess spending for all delayed products varied widely based on the rebate level used (exhibit 3). In the second sensitivity analysis, in which we assumed no discount for generic products in the first 180 days of product entry, the excess costs were similar: $743 million compared to $761 million in the primary analysis.

SOURCE Authors’ analysis of data for 2010–16 from the Medicaid State Drug Utilization files. NOTE The exhibit shows the results from our first sensitivity analysis, in which we varied the brand-name rebates from 0 percent to 90 percent.

Causes For Delay In Generic Entry

Appendix exhibit A3 outlines the reasons for delays in generic entry for the top fifteen brand-name products by total Medicaid spending.¹² For cases in which delayed generic entry was observed, there was often more than one reason associated with the delay. For twelve (80 percent) of these products, patent litigation on later-issued patents was the most common reason associated with delays in generic entry; nine of the twelve examples in which litigation was a contributor were settled out of court. Ten of the fifteen products (67 percent) also had a citizen petition filed that might have delayed FDA approval of the generic drug. In the case of bosentan (Tracleer), the generic manufacturers cited issues with obtaining brand-name samples for bioequivalence testing. Patent litigation was also a common theme among the top five products: The brand-name manufacturers of darunavir and olopatadine successfully defended their patents in court (after doing so, the brand-name manufacturer for darunavir also settled out of court and reached licensing agreements with three generic manufacturers). Disputes over patents related to atorvastatin and pioglitazone were settled out of court after delaying generic entry by approximately 1.75 and 1.50 years, respectively.

Discussion

Amid growing concern relating to US pharmaceutical spending, we surveyed the prevalence of delays in generic drug entry and the costs of such delays to the health care system. Using Medicaid data for 2010–16, we found that for more than half of the products in our cohort, generic entry occurred within a quarter-year of the expected date (either earlier than expected or as expected). When entry was delayed, it was often delayed for more than a year. The excess costs associated with delays in generic entry incurred by Medicaid alone were estimated to be in the hundreds of millions of dollars over the study period, with the top five drugs in terms of total Medicaid spending contributing nearly two-thirds of this amount. In the subgroup of products for which generic entry was delayed, litigation over later-issued patents was identified as the most common reason for delay.

A prior study that used Medicaid data for 2000–10 and used the same method we did to approximate generic entry found that for 64 percent of brand-name products in the sample, generic entry occurred before or within one quarter-year of the expected date and was delayed for the other 36 percent.⁶ Unlike this study, though, it did not account for products that did not experience generic entry. We estimated the pre- and post-rebate excess spending due to delays in generic entry to be $3.3 billion and $761 million, respectively. However, because Medicaid serves only 23 percent of the total US insured population,²⁰ and non-Medicaid plans (that is, commercial insurance and Medicare plans) typically have much lower brand-name rebates, excess spending in other parts of the health care system are likely much higher.

Although prior literature has identified several strategies that could delay generic entry,¹⁰ our study identified patent litigation as the most common reason for delay, affecting 80 percent of the top fifteen drugs by total Medicaid spending in our cohort. When patent litigation was observed, most cases were settled out of court, which raises concerns that brand-name manufacturers may be offering incentives to generic manufacturers in exchange for terminating ongoing patent challenges. Such cases should be reviewed by the Federal Trade Commission (FTC) to determine their legality, since the FTC has in the past been particularly concerned about settlements that involve a cash payment from the brand-name manufacturer to the generic manufacturer in exchange for delayed market entry, or “pay-for-delay” settlements. Filing citizen petitions with the FDA was also common, although the impact of this strategy on time to generic entry was less clear. Other widely discussed strategies, such as restricting access to drug samples for bioequivalence testing, played an obvious role in only a few cases.²¹

This study leads to certain policy recommendations. First, we were able to use patent and regulatory filing information to predict expected generic entry. Predicted generic entry dates could be used to inform regulators and payers of a delay in generic entry and lead to further investigation of the cause of that delay. Such investigations could then lead to appropriate remedial measures to provide timely access to generic products for patients.

Second, our study highlights the role of additional patents and litigation on the delays in market entry of generic products. Although these later-issued patents are more likely to be invalidated during litigation than the key patents on the product’s active ingredient,^22,23 patent litigation can be costly and time-consuming to all parties involved.²⁴ To avoid these expenses, the USPTO could establish more rigorous standards when reviewing and issuing drug patents.²⁵ A significantly less costly and less used option that could expedite the resolution of patent disputes is a process known as inter partes review, an administrative process at the USPTO where disputes are required to be resolved within a year.²⁶Inter partes review could be required of all new patents when they are filed with the FDA, to weed out improperly granted patents before they require extensive litigation in court.

Conclusion

Our findings underscore the financial importance of mitigating delays in generic drug entry and highlight patent litigation as a key contributor to these delays. Policies that effectively balance incentives for drug development against the public good associated with robust generic drug competition and use should be implemented. In the near term, addressing inefficiencies resulting from patent litigation should be considered, since the costs of delayed generic entry to the US health care system can be substantial.

ACKNOWLEDGMENTS

Research was supported by the PCMA Foundation, Arnold Ventures, and the Harvard-MIT Center for Regulatory Science.

NOTES

• 1 Nominal drug spending hit $452.6 billion in the US in 2017, says IQVIA. Pharmaceutical Commerce [serial on the Internet]. 2018 Apr 22 [cited 2020 Feb 24]. Available from: https://pharmaceuticalcommerce.com/brand-marketing-communications/nominal-drug-spending-hit-452-6-billion-in-the-us-in-2017-says-iqvia/ Google Scholar
• 2 Kesselheim AS, Avorn J, Sarpatwari A. The high cost of prescription drugs in the United States: origins and prospects for reform. JAMA. 2016;316(8):858–71. Crossref, Medline, Google Scholar
• 3 Wang B, Liu J, Kesselheim AS. Variations in time of market exclusivity among top-selling prescription drugs in the United States. JAMA Intern Med. 2015;175(4):635–7. Crossref, Medline, Google Scholar
• 4 Kesselheim AS, Sinha MS, Avorn J. Determinants of market exclusivity for prescription drugs in the United States. JAMA Intern Med. 2017;177(11):1658–64. Crossref, Medline, Google Scholar
• 5 Kesselheim AS. An empirical review of major legislation affecting drug development: past experiences, effects, and unintended consequences. Milbank Q. 2011;89(3):450–502. Crossref, Medline, Google Scholar
• 6 Beall RF, Darrow JJ, Kesselheim AS. A method for approximating future entry of generic drugs. Value Health. 2018;21(12):1382–9. Crossref, Medline, Google Scholar
• 7 Beall RF, Darrow JJ, Kesselheim AS. Patent term restoration for top-selling drugs in the United States. Drug Discov Today. 2019;24(1):20–5. Crossref, Medline, Google Scholar
• 8 Sinha MS, Najafzadeh M, Rajasingh EK, Love J, Kesselheim AS. Labeling changes and costs for clinical trials performed under the US Food and Drug Administration pediatric exclusivity extension, 2007 to 2012. JAMA Intern Med. 2018;178(11):1458–66. Crossref, Medline, Google Scholar
• 9 Kapczynski A, Park C, Sampat B. Polymorphs and prodrugs and salts (oh my!): an empirical analysis of “secondary” pharmaceutical patents. PLoS One. 2012;7(12):e49470. Crossref, Medline, Google Scholar
• 10 Vokinger KN, Kesselheim AS, Avorn J, Sarpatwari A. Strategies that delay market entry of generic drugs. JAMA Intern Med. 2017;177(11):1665–9. Crossref, Medline, Google Scholar
• 11 Medicaid.gov. State Drug Utilization Data (SDUD) FAQs [Internet]. Baltimore (MD): Centers for Medicare and Medicaid Services; [cited 2019 Dec 10]. Available from: https://www.medicaid.gov/medicaid/prescription-drugs/state-drug-utilization-data/state-drug-utilization-data-faq/index.html Google Scholar
• 12 To access the appendix, click on the Details tab of the article online.
• 13 IMS Institute for Healthcare Informatics. Price declines after branded medicines lose exclusivity in the U.S. [Internet]. Parsippany (NJ): The Institute; 2016 Jan [cited 2019 Dec 10]. Available from: https://www.iqvia.com/-/media/iqvia/pdfs/institute-reports/price-declines-after-branded-medicines-lose-exclusivity-in-the-us.pdf Google Scholar
• 14 Medicaid and CHIP Payment and Access Commission. MACStats: Medicaid and CHIP data book [Internet]. Washington (DC): MACPAC; 2018 Dec [cited 2019 Dec 10]. Available from: https://www.macpac.gov/wp-content/uploads/2018/12/December-2018-MACStats-Data-Book.pdf Google Scholar
• 15 Korn DE, Lietzan E, Scott SW. A new history and discussion of 180-day exclusivity. Food Drug Law J. 2009;64(2):335–90. Medline, Google Scholar
• 16 Dickson S. Effect of evergreened reformulations on Medicaid expenditures and patient access from 2008 to 2016. J Manag Care Spec Pharm. 2019;25(7):780–92. Crossref, Medline, Google Scholar
• 17 Dave CV, Hartzema A, Kesselheim AS. Prices of generic drugs associated with numbers of manufacturers. N Engl J Med. 2017;377(26):2597–8. Crossref, Medline, Google Scholar
• 18 Dave CV, Kesselheim AS, Fox ER, Qiu P, Hartzema A. High generic drug prices and market competition: a retrospective cohort study. Ann Intern Med. 2017;167(3):145–51. Crossref, Medline, Google Scholar
• 19 Dave CV, Pawar A, Fox ER, Brill G, Kesselheim AS. Predictors of drug shortages and association with generic drug prices: a retrospective cohort study. Value Health. 2018;21(11):1286–90. Crossref, Medline, Google Scholar
• 20 Henry J. Kaiser Family Foundation. Health insurance coverage of the total population, 2017 [Internet]. San Francisco (CA): KFF; [cited 2020 Apr 15]. Available from: https://www.kff.org/other/state-indicator/total-population/?currentTimeframe=0&sortModel=%7B%22colId%22:%22Location%22,%22sort%22:%22asc%22%7D Google Scholar
• 21 Brill A. Lost prescription drug savings from use of REMS programs to delay generic market entry [Internet]. Washington (DC): Matrix Global Advisors; 2014 Jul [2019 Dec 10]. Available from: https://static1.1.sqspcdn.com/static/f/460582/25228342/1406034596510/REMS_Study_July.pdf Google Scholar
• 22 Hemphill CS, Sampat BN. Weak patents are a weak deterrent: patent portfolios, the Orange Book listing standard, and generic entry in pharmaceuticals [Internet]. Paper presented at: Standards, Patents, and Innovations (NBER conference), 2012 Jan 20–21; [cited 2020 Apr 15]. Available from: https://conference.nber.org/conferences/2012/IPKE/sampat.pdf Google Scholar
• 23 Hemphill CS, Sampat BN. When do generics challenge drug patents? J Empir Leg Stud. 2011;8(4):613–49. Crossref, Google Scholar
• 24 American Intellectual Property Law Association. 2015 report of the Economic Survey. Arlington (VA): AIPLA; 2015 Aug. Google Scholar
• 25 Furrow ME. Pharmaceutical patent life-cycle management after KSR v. Teleflex. Food Drug Law J. 2008;63(1):275–320. Medline, Google Scholar
• 26 Cohen EC. A primer on inter partes review, covered business method review, and post-grant review before the Patent Trial and Appeal Board. Fed Cir Bar J. 2014;24(1):1–34. Google Scholar