Trends in the regulatory review period

Transcription

1 Tr ends Trends Are Development Times For Pharmaceuticals Increasing Or Decreasing? Thefindingsofthisstudysuggestthatdevelopmenttimesarenota factor in rising drug prices. by Salomeh Keyhani, Marie Diener-West, and Neil Powe ABSTRACT: This study examines trends in drug development times. Longer clinical trial times have been described as one factor leading to higher drug prices. Previous reports on development times have been based on proprietary data. We examined trends in development times for 168 drugs with data collected from publicly available sources. The median clinical trial and regulatory review periods for drugs approved between 1992 and 2002 were and years, respectively. Clinical trial periods have not increased during this time frame, and regulatory review periods have decreased. Therefore, it is unlikely that longer clinical trial times are contributing to rising prescription drug prices. [Health Affairs 25, no. 2 (2006): ; /hlthaff.2.461] Trends in the regulatory review period of drugs over time have been extensively examined, because public data are available for this phase. 1 However, information on clinical trial periods is mainly based on proprietary data supplied by the pharmaceutical industry that are unavailable to the public, or on duration models based on data supplied by industry news and intelligence services. This prior work suggests that clinical trial time has been decreasing. 2 Our examination of development time adds to the earlier work in the area; we were able to reveal data on individual drugs, analyze development times by various drug characteristics, and examine trends in ways that have not been published before, mainly because we were not limited by proprietary agreements with industry. In addition, we used government-verified development dates on approved drugs, not proprietary data, to examine trends. Development time is important because of numerous claims that it is a factor in drug pricing. The pharmaceutical industry argues that long drug development times lead to high drug prices, and it identifies increasingly longer clinical trial periods as an important factor. 3 Long development times presumably lead to higher opportunity costs of capital and higher research and development (R&D) costs. It has also been suggested that decreases in development time can have a substantial downward effect on these costs. 4 The claim that longer development times lead to high drug prices suggests that drug pricing is based on costs. In other words, high R&D costs lead to higher drug prices, and current prices are necessary to fund future R&D spending. Salomeh Keyhani (Salomeh.Keyhani@mountsinai.org) is an assistant professor of health policy at Mount Sinai School of Medicine in New York City. Marie Diener-West is a professor of biostatistics, ophthalmology, and epidemiology at the Johns Hopkins Bloomberg School of Public Health in Baltimore, Maryland. Neil Powe is a professor of medicine, epidemiology, and health policy and management at the Bloomberg School. HEALTH AFFAIRS ~ Volume 25, Number DOI /hlthaff Project HOPE The People-to-People Health Foundation, Inc.

2 Health Tracking Market factors probably play a much larger role than development times in determining drug prices. 5 Previous studies have found that therapeutic value and market structure are the main explanatory variables for pricing of new molecular entities (NMEs). For example, John Lu and William Comanor found that introductory prices for drugs that represented an important therapeutic gain were two to three timeshigherthanpricesfordrugsthatwere used for similar purposes. 6 However, given the claim that longer development times lead to higher drug prices and the close attention paid by policymakers to drugs affordability, a closer look at development times is warranted. Development time is also of interest as it has an indirect effect on patent life. A shorter development period can lead to longer patent duration, because patent applications are usually filed prior to drug development. In previously published work we compared the prices of different groups of drugs after accounting for development time, government support, market size, and other drug characteristics. 7 In this paper we focus solely on development time and analyze trends based on publicly available data. Study Data And Methods We conducted a retrospective study examiningdevelopmenttimesbydrugcharacteristics and government regulatory designations. To be eligible for this study, a human therapeutic drug (we excluded development data on nonhuman drugs and diagnostic medications) hadtobelistedinthefederal Register Government Printing Office (GPO) Web site before May 2003 (the date that data collection was completed).thisistheonlypubliclyavailable source of data on drug development time. 8 For purposes of comparison, we also collected data on review times from the U.S. Food and Drug Administration (FDA) Web site. In addition, we collected data on a variety of factors thought to influence development time: FDA regulatory designations, because drugs that have received fast-track or accelerated review might be rapidly developed; use category (acute, chronic, intermittent), because clinical trial times might vary for drugs that are used for different periods; orphan drug status, because such drugs target rare diseases and might take longer to develop; first-in-class status, because drugs that are novel might take longer to develop; and annual sales, because drugs with higher potential sales might be developed more rapidly. Development process. The drug development process involves three major stages: pre clinical phase, clinical trial phase, and regulatory review phase. 9 When a compound with potential clinical benefit is identified, a drug company files an investigational new drug (IND) application with the FDA. The IND application is a compilation of all known information about the compound. If the FDA does not intervene after thirty days, the IND becomes effective, and the clinical trial period can begin. The clinical trial period involves testing of a new compound in humans. The regulatory review period starts with the filing of a new drug application (NDA) with the FDA for marketing a new drug and ends with approval of the drug. There are no publicly available data for the first phase (before an IND application). Under the Drug Price Competition and Patent Term Restoration Act of 1984, drug companies have the option of filing for patent term restoration with the U.S. Patent and Trademark Office (USPTO), which can result in the retrospective addition of the length of the clinical trial and regulatory review phase to the patent term. 10 This option is available for all drugs that have less than fourteen years of patent life remaining at the time of approval. When filing, companies are required to declare both the IND and NDA dates. The FDA verifies the dates and publishes a notice in the Federal Register. This notice provides the dates necessary to calculate the clinical trial and regulatory review periods of a particular drug. For our study, the clinical trial period is the time between the date the IND becomes effective and the date the NDA is filed. The regulatory review period starts with filing of the NDA and ends with the approval of the drug. Drug characteristics. We defined use 462 March/April 2006

3 Tr ends categories based on the information listed in the FDA-approved drug label. Acute drugs are those used fewer than thirty consecutive days by patients; chronic drugs are those used more than thirty days; and intermittent drugs are used once, once a week, or once every few weeks. When the drug label did not give details on how long a drug should be used, we based the categorization on how long the drughadbeenusedintheclinicaltrialsdescribed. We collected data on orphan drugs, a regulatory designation for drugs used to treat rare diseases, from the FDA Web site. First-in-class was defined as a drug first in its chemical class to be approved for its indication. We obtained 2002 annual sales data from IMS Health. We defined a drug as a biopharmaceutical if it was approved by the FDA Center for Biologic Evaluation and Research (CBER) rather than the FDA Center for Drug Evaluation and Research (CDER). We obtained data on fast-track status and accelerated review from the FDA Web site. Fasttrack status was described in the 1997 FDA Modernization Act as a formal mechanism that allows industry to interact with the FDA and that includes a number of procedures that can lead to more rapid drug approval. 11 Accelerated review is a program intended to allow products for life-threatening diseases to be available on the market more quickly on the basis of preliminary evidence before patient benefit has been formally demonstrated. Statistical analysis. We used nonparametric statistical methods to examine differences in the distribution of total development times among the various drug categories. Differences between two groups were assessed using the Wilcoxon rank-sum test; differences amongthreeormoregroupswereassessedusing the Kruskal-Wallis test. All analyses were performed using Stata 8.0. Study Findings Sample development times. A total of 168 human therapeutic drugs approved between 1 January 1992 and 1 January 2002 that met the eligibility criteria were identified in the Federal Register. A total of 298 human therapeutic new molecular entities (NMEs) were approved by the FDA in this period. Our study included 145 (approximately half) of these NMEs and another twenty-three drugs that are not NMEs (mainly biologics and vaccines). Exhibit 1 shows the twenty-five drugs with the highest annual sales (2002) in our sample. The median total post-ind development time foralldrugsinthedatabasewas6.4years;it ranged from 2.6 to 17.3 years (data not shown). The median clinical trial phase was years and ranged from 1.4 years to 14.6 years, while the median regulatory review phase was years and ranged from 0.3 years to 5.0 years. Development time by drug characteristics. Overall, the median total post-ind development time of biopharmaceuticals, drugs first in their chemical class, and orphan drugs appears to be longer than that of their counterparts; however, these differences were not statistically significant (Exhibit 2). It appears that drugs used for acute indications (pneumonia), for example, have shorter clinical trial phases than drugs used chronically (such as for depression or dementia). However, these differences do not appear to be statistically significant. The median development time for fast-track drugs was more than two years shorter than that for drugs without this regulatory designation. Interestingly, drugs that had annual sales greater than $100 million took one year less to develop than drugs with lower annual sales. Trends in development time. Development time trends can be examined in two ways: by year of NDA approval or by year of IND filing. Overall, it appears that the total post-ind development and regulatory review periods have decreased over time (Exhibit 3). Clinical trial periods have not increased and appear to be trending downward in recent years. The median regulatory review period of all drugs approved each year (available on the FDA Web site) was similar to that for drugs obtained from the Federal Register Web site (Exhibit 3), which suggests that the sample of drugsisrepresentativeofdrugsingeneral. Exhibit 4 shows the median development time for a group of drugs for which develop- HEALTH AFFAIRS ~ Volume 25, Number 2 463

4 Health Tracking EXHIBIT 1 Sample Of Development Times For Twenty-Five Drugs With The Highest Annual Sales In The Federal Register Brand name Common treatment indication(s) IND date a NDA filed NDA approved Clinical trial period (years) Regulatory review period (years) Lipitor Prevacid Neurontin Nexium Allegra High cholesterol Reflux, ulcers Seizures Reflux, ulcers Allergies 10/31/90 07/03/87 07/02/86 08/18/97 11/04/93 06/17/96 11/15/93 01/31/92 12/03/99 07/31/95 12/17/96 05/10/95 12/30/93 02/20/01 07/25/ Risperdal Effexor XR Fosamax Plavix Celexa Psychotic disorder/mania Depression Osteoporosis Stroke Depression 09/07/88 11/23/85 09/29/88 05/05/90 07/30/83 04/15/92 04/26/91 03/31/95 04/28/97 05/07/97 12/29/93 12/28/93 09/29/95 11/17/97 07/17/ Actos Zyrtec Singulair Levaquin Avandia Diabetes Allergies Asthma/allergies Pneumonia/infections Diabetes 10/21/89 05/31/84 06/03/92 05/04/91 12/23/93 01/15/99 07/01/88 02/21/97 12/21/95 11/24/98 07/15/99 12/08/95 02/20/98 12/20/96 05/25/ Protonix Seroquel Viagra Aciphex Enbrel Reflux/ulcer Schizophrenia/mania Impotence Reflux/ulcer Rheumatoid arthritis 10/13/90 10/21/88 01/08/95 08/21/91 06/26/92 06/30/98 07/29/96 09/29/97 03/31/98 05/08/98 02/20/00 09/26/97 03/27/98 08/19/99 11/02/ Taxotere Aricept Topamax Evista Detrol Breast cancer Dementia Seizures Osteoporosis Overactive bladder 11/18/90 01/23/91 06/18/86 02/16/83 10/07/94 07/27/94 04/05/96 08/08/94 06/09/97 03/25/97 05/14/96 11/25/96 12/24/96 12/09/97 03/25/ SOURCE: Drug development data obtained from the Federal Register U.S. Government Printing Office Web site. NOTES: Drugs listed in order of highest annual sales. NDA is new drug application. The median clinical trial period for all drugs is years; the median regulatory review period for all drugs is years. a Investigational new drug (IND) date refers to the date that clinical trials can be initiated. This date is usually thirty days after an IND is filed with the U.S. Food and Drug Administration (FDA). ment was begun in the same year. Each cohort was under the same FDA regulatory framework at the time of filing the IND. The median total post-ind development time (clinical trial period plus regulatory review period) for combined drugs has decreased between 1985 and Available data on the small number of drugs with INDs filed before 1985 and after 1995 support the same trend (data not shown). Study Limitations Potential for selection bias. Anumber of limitations of this study deserve comment. Our data set is composed of drugs for which patent extensions were filed. The reasons for filing or not filing a patent extension could lead to selection bias. Industry might not file an extension with the USPTO for a number of reasons. 12 For example, if the marketing applicant is different from the patent owner, an application may not be filed. The manufacturer might miss the application period for patent term restoration, because the application must be filed with the patent office within sixty days of the drug s approval. Another possible limitation is that industry factors might affect the decision to file for patent extension, which could in turn bias the sample and subsequently the development trends observed. In other words, industry could weigh the economic gain versus the cost of filing for an extension. However, it is unlikely that industry would forgo increased patent protection on any drug simply because the costs are small 464 March/April 2006

5 Tr ends EXHIBIT 2 Pharmaceutical Development Times, By Category Post-IND development a Clinical trial period (years) Regulatory review period (years) Category Median Range p value b Median Range Median Range Biopharmaceutical First in class Annual sales >$100m c Use category Acute Chronic Intermittent Route Oral Intravenous Topical FDA review Fast track Accelerated review Orphan drug SOURCE: Drug development data obtained from the Federal Register U.S. Government Printing Office Web site. NOTE: FDA is Food and Drug Administration. a Post-IND (investigational new drug) development time includes both the clinical trial period and regulatory review period. b P value less than.05 suggests statistically significant differences. c $100 million was chosen as a cutoff point because roughly half the drugs in the database had sales greater than this figure and half had sales below this figure. compared with the potential revenue gained from increased patent life. The USPTO patent extension fee is $1,120. The associated attorney fees that might accompany such an application probably exceed the USPTO fees; however, thesecostsarestilllikelytobenegligiblecomparedwiththeincreasedrevenueandcompetitive advantage that might be gained from an increased patent term. 13 Given the importance of intellectual property rights to the economic health of this industry, it is doubtful that any increased patent life would be overlooked. Finally, a drug is not eligible for patent extension if more than fourteen years of patent life remain when the drug is approved. The logical extension of this point is that our sample represents a group of drugs that on average took longer to develop, and we might be overestimating drug development times. However, there is no other source of public data on clinical trial periods. In addition, comparisons of our data with other public data on regulatory review periods (FDA Web site) suggest that our sample is representative (Exhibit 3). Pros and cons of analytic method. We examined trends in development time by the HEALTH AFFAIRS ~ Volume 25, Number 2 465

6 Health Tracking EXHIBIT 3 Total Post-IND Development Period, Clinical Trial Period, And Regulatory Review Periods, By Year Of NDA Approval, Median development time (years) 8 6 Total post-ind development period Clinical trial period Regulatory review period FDA-reported regulatory review period Year of NDA approval SOURCE: Drug development data obtained from the Federal Register U.S. Government Printing Office Web site and the U.S. Food and Drug Administration (FDA) Web site. NOTES: The FDA-reported regulatory review period is based on public data on review periods available on the FDA Web site for all drugs approved by the FDA Center for Drug Evaluation and Research (CDER) each year. IND is investigational new drug. NDA is new drug application. year in which the IND was filed as well as by the year in which the NDA was approved. Each method has its limitations. Exhibit 3 demonstrates the changes in median development time charted by the year NDAs are approved. Given that the initiation of each drug s development started during a different year, a clear picture of how drug development time has changed over time does not emerge. Exhibit 4 demonstrates the change in median development time charted by the year of IND filing. The effects of legislation on development time EXHIBIT 4 Total Post-IND Development Period, Clinical Trial Period, And Regulatory Review Period, By Year Of IND Filing, Median development time (years) 10 8 Total post-ind development period Clinical trial period Regulatory review period Year of IND filing 1995 SOURCE: Drug development data obtained from the Federal Register U.S. Government Printing Office Web site and the U.S. Food and Drug Administration (FDA) Web site. NOTES: The FDA-reported regulatory review period is based on public data on review periods available on the FDA Web site for all drugs approved by the FDA Center for Drug Evaluation and Research (CDER) each year. IND is investigational new drug. 466 March/April 2006

7 Tr ends can be more clearly examined, because the development of each cohort of drugs was initiated in the same year. However, because we do not have data for all drugs for which INDs were filed every year, we might be focusing on drugs that took much less time to develop. For example, there still could be drugs that are now in development for which INDs were filedin1992.however,oursampleincludes drugs approved up to 2002, and the trend is present during the entire period. Concluding Comments Our findings indicate that drug development times vary by drug characteristics and have decreased over time. There is some suggestionthatdrugsthattreatrarediseases(orphan drugs) and first-in-class drugs take longer to develop. One would expect it to be harder, for example, to find and enroll patients with rare diseases into clinical trials. Drugs that are first-in-class might be harder to develop because these drugs are being tested by industry and reviewed by the FDA for the first time. This could explain the differences we found in development times, although they were not statistically significant. Drugs with higher annual sales appear to be developed more quickly. This could represent a company s financial interest in more rapidly developing a drug that appears to be a blockbuster. It also appears that FDA regulatory designations such as fast-track status have resulted in decreased development times; however, fasttrack status was not implemented until 1998, and if the development times of drugs that received fast-track status are compared to drugs that were approved after 1998, the difference in median development time is only 1.4 years (asopposedto2.3yearsinexhibit2).this suggests that development time overall has been decreasing independently of the effect of such regulatory designations. Thefactthatdevelopmenttimeshavebeen decreasing for each new cohort of drugs with INDs filed in the same year is especially informative. Design, recruitment, and implementation of clinical trials by industry might be becoming more efficient. In addition, the FDA might have become better in communicating its expectations on clinical trial design and quality to industry. Longer clinical trial periods have been offered as a justification for increasing drug prices; our analysis demonstrates that these periods have not been increasing and might in fact be decreasing. This finding, in turn, suggests that development times are not a factor in rising drug prices. Previous research has demonstrated that drug prices are affected by marketfactorssuchastheavailabilityofsubstitutes and the relative therapeutic advantage of one drug over another that treats similar diseases. 14 The argument that long drug development times are a reason for high drug prices might be more politically acceptable than the explanation that certain market factors determine price and that firms will price new drugs to maximize profits. The 1984 Hatch-Waxman act allowed the pharmaceutical industry to add the period that drugs have been under testing and review to the length of the patent. Recent reports suggest that patent life has been increasing because of new patent provisions offered as new incentives to industry for drug discovery and development. 15 Given the increase in patent life reported elsewhere and the decrease in drug development times noted in this study, incentives that lead to increased patent life need to be examined carefully. In conclusion, clinical trial times are not increasing and are an unlikely contributor to rising prescription drug prices. Salomeh Keyhani is a former Robert Wood Johnson Clinical Scholar, supported by the Robert Wood Johnson Foundation. This research is the sole responsibility of the authors. They acknowledge the detailed and thoughtful input of the anonymous reviewers who helped improve this manuscript. HEALTH AFFAIRS ~ Volume 25, Number 2 467

8 Health Tracking NOTES 1. M.K. Olson, Managing Delegation in the FDA: Reducing Delay in New-Drug Review, Journal of Health Politics, Policy and Law 29, no. 3 (2004): ; D.P. Carpenter, The Political Economy of FDADrugReview:Processing,Politics,andLessons for Policy, Health Affairs 23, no. 1 (2004): 52 63; and D. Carpenter et al., Approval Times for New Drugs: Does the Source of Funding for FDA Staff Matter? Health Affairs 22 (2003): w618 w624 (published online 17 December 2003; /hlthaff.w3.618); and M.K. Olson, Explaining Reductions in FDA Drug Review Times: PDUFA Matters, Health Affairs 23 (2004): s1 s2 (published online 30 January 2004; / hlthaff.w4.s1). 2. J.M. Reichert, Trends in Development and Approval Times for New Therapeutics in the United States, Nature Reviews: Drug Discovery 2, no. 9 (2003): ; K.I. Kaitin et al., Measuring thepaceofnewdrugdevelopmentintheuser Fee Era, Drug Information Journal 34, no. 3 (2000): ; and R.M. Abrantes-Metz, C.P. Adams, and A. Metz, Pharmaceutical Development Phases: A Duration Analysis, Working Paper no. 274 (Washington: Bureau of Economics, Federal Trade Commission, 2004). 3. Pharmaceutical Research and Manufacturers of America, Why Do Prescription Drugs Cost So Much and Other Questions about Your Medicines, June 2000, publications/publications/brochure/questions (accessed 31 August 2005); and PhRMA, The Search for a Cure for Cancer: New Medicines, New Hope, Summer 2003, (accessed 28 January 2005). 4. J.A.DiMasi,R.W.Hansen,andH.G.Grabowski, The Price of Innovation: New Estimates of Drug Development Costs, Journal of Health Economics 22, no. 2 (2003): ; and J.A. DiMasi, The Value of Improving the Productivity of the Drug Development Process: Faster Times and Better Decisions, PharmacoEconomics 20, no. 3 Supp. (2002): S.O. Schweitzer, Pharmaceutical Economics and Policy (New York: Oxford University Press, 1997). 6. Z.J. Lu and W.S. Comanor, Strategic Pricing of New Pharmaceuticals, Review of Economics and Statistics 80, no. 1 (1998): S. Keyhani, M. Diener-West, and N. Powe, Do Drug Prices Reflect Development Time and Government Investment? Medical Care 43, no. 8 (2005): This paper includes data on twelve additional antiretroviral drugs for which development data were obtained from sources other than the Federal Register. We chose to include data on antiretroviral drugs because the purpose of our previous study was to examine differences in prices between different drug classes after accounting for development time and other factors. Data on development time other than those available in the Federal Register are scarce; however, given interest in AIDS drugs early in the epidemic, data on these drugs were occasionally revealed by industry. We chose not to include these drugs in this study because they represent a highly select group that had rapid developmentandmighthavebiasedourdataon development time downward. 8. The Federal Register ( fr/advanced.html) can be searched to find notices related to patent term restoration. 9. U.S. Congress Office of Technology Assessment, Pharmaceutical R&D: Costs, Risks, and Rewards, February 1993, disk1/1993/9336_n.html (accessed 18 April 2005). 10. U.S. Food and Drug Administration, Frequently Asked Questions on the Patent Term Restoration Program, 8 December 2003, (accessed 29 vember 2005). 11. FDA, Fast Track, Priority Review, and Accelerated Approval, 31 January 2006, (accessed 31 January 2006). 12. FDA, Frequently Asked Questions. 13. U.S. Patent and Trademark Office, Formal Requirements for Application for Extension of Patent Term; Correction of Informalities Patent Rules, mpep/documents/appxr_1_740.htm (accessed 31 January 2006); and FDA, Dockets Management, (accessed 31 January 2006), for an actual application. 14. Lu and Comanor, Strategic Pricing. 15. H.Soehnge, TheDrugPriceCompetitionand Patent Term Restoration Act of 1984: Fine-Tuning the Balance between the Interests of Pioneer and Generic Drug Manufacturers, Food and Drug Law Journa1 58, no. 1 (2003): 51 80; and National Institute for Health Care Management Research and Educational Foundation, Prescription Drugs and Intellectual Property Protection: Finding the Right Balance between Access and Innovation, August 2000, (accessed 12 January 2005). 468 March/April 2006