Drug launch curves in the modern era, a new article in the January 2017 issue of Nature Reviews: Drug Discovery, contains projections of relevance for forecasting pharmaceutical sales in the US market. Relying on unit sales for 61 innovative drugs receiving FDA approval between 2000 and 2002, the authors determined that the median product follows an S-shaped launch-to-peak penetration curve: achieving 11% of peak sales in Year 1, 31% in Year 2, 58% in Year 3, 76% in Year 4, 89% in Year 5, and 100% (i.e. peak sales) in Year 6. On an interquartile basis, time-to-peak (TTP) ranged from 4 to 9 years; the minimum TTP was 2 years and the maximum was 14.
Using data from the Supplemental Materials file we find that the average for TTP for first-in-class drugs is 5.65 years as opposed to 6.93 years for subsequent entrants. Unlike other research on the subject, however, the authors did not that there was a statistically significant relationship between entry order and TTP. But, as they note, TTP is distinct from factors such as market share and revenue potential that may be more indicative of first-mover advantage.
Anyone interested in drug development will want to take a look at Clinical Development Success Rates 2006-2015, a new report from the Biotechnology Innovation Organization (BIO) in partnership with Amplion and Biomedtracker. Per the associated press release, the study “recorded and analyzed 9,985 clinical and regulatory phase transitions, across 1,103 companies” over the past decade to calculate phase-success rates (i.e. the probability of moving forward from Phase 1 to Phase 2, from Phase 2 to Phase 3, etc.) as well as the overall likelihood of approval (LOA).
The results of the BIO study are similar to those reported by Joseph DiMasi and others at the Tufts Center for the Study of Drug Development (CSDD) who looked at 1,442 compounds from 50 leading pharmaceutical firms that initially entered clinical testing between 1995 and 2007. BIO found slightly higher Phase 1 to Phase 2 transition rates but somewhat lower success rates for each step thereafter, leading to an average LOA of 9.6% versus 11.8% as reported by the CSDD. It’s interesting to contemplate whether the higher number of investigational agents moving from Phase 1 to Phase 2 in the BIO study is the reason for lower success rates in subsequent phases. If that is the case, it suggests that the industry needs to work on “failing fast” so that resources are not unnecessarily expended on unpromising candidates.
In addition to aggregate success rates, the BIO report provides detailed figures by disease state and finds that therapies aimed at hematologic or infectious conditions have the highest LOAs (26.1% and 19.1%, respectively) while oncologic and psychiatric agents have the lowest LOAs (5.1% and 6.2%).
Relative success rates for drugs aimed at rare diseases versus those aimed at chronic, high-prevalence conditions were also examined. It turns out that rare drug development programs have a 25.3% LOA as compared to 8.7% for the chronic condition set. The industry shift towards orphan drug development over the past decade is well documented and is frequently attributed to the high prices commanded by therapies for small-population diseases. In forecasting models, however, a three-fold greater likelihood of actually making it to market is almost certainly going to be a bigger driver of net present value than price. The emphasis on orphan drugs has, of course, arisen from a very complex set of factors but improved approval odds are hard to ignore; having these odds quantified in a robust data analysis may well accelerate the flow of development funds towards rare diseases.
The BIO report is free and can be downloaded here.
The DiMasi / Tufts figures are reported in the May 2016 issue of the Journal of Health Economics.
As we all know, executives from both Turing Pharmaceuticals and Valeant Pharmaceuticals have been brought before Congressional panels to address questions arising from the furor over their respective pricing tactics. Last week, the Senate Special Committee on Aging released an 818-page PDF file containing testimony associated with Valeant’s appearance on April 27.
Although it seems like the basic facts are pretty clear at this point, industry critics and supporters will probably find material to reinforce their arguments among the documents. On a separate note, they provide opportunity to review information that is very rarely made public. Pharmaceutical pricing in the US is frequently opaque and almost always very complicated. Due to confidentiality agreements, even people who are actively involved in pricing decisions rarely get to see the details of what other firms are doing. Leaving aside the top-line controversy over price increases, these documents allow for an unexpected “peak under the hood” at the mechanics of another company’s process.
A recent Wall Street Journal article addresses the fact that biosimilars won’t be replicating the savings generated by small molecule generics anytime soon. None of this should come as a surprise.
Manufacturing biologics is substantially more complicated and more expensive than is the case for small molecule drugs. Often, the production process has to be tailor-made for each product: what works for one won’t work for another. Adding another layer of complexity is the fact that biologic manufacturing processes are themselves protected by numerous forms of intellectual property and trade secrets. To replicate the characteristics of the final product, biosimilar manufacturers must, to a greater or lesser extent, re-create the production process using other means. So, while a biosimilar company can avoid a good portion of the R&D costs, it still has to create what amounts to proprietary manufacturing methods.
Not that biosimilar hopefuls are completely off the hook for R&D either. Bioequivalence studies don’t satisfy the FDA’s current thinking of what it takes to approve these copycat products for use, where the standard is that “there are no clinically meaningful differences between the biological product and the reference product in terms of the safety, purity, and potency of the product.” Demonstrating equivalence for biosimilars may necessitate data from new clinical studies.
All this is illustrated by a 2013 article from the National Library of Medicine and a study by the IGES Institute that estimated developing a biosimilar takes an average of seven to eight years and costs $100 million to $250 million, while bringing traditional generics to market only costs between $1 million and $4 million.
Then there is the matter of who is allowed to replace the brand product with the biosimilar. With generics, the doctor or pharmacist can make the switch and, at present, this is almost always done automatically via IT systems at the pharmacy or other dispensing location. Substation rights are another hotly contested area with biosimilars and, since this is a regulatory area that is largely governed at the state level, the result is an emerging patchwork of inconsistent rules across the country – many creating barriers to getting a biosimilar to the patient. Automatic substitution creates a guaranteed demand for generics that, for the foreseeable future at least, will not be present for biosimilars.
In this environment, modest discounts are to be expected. Classical economic theory tells us that price declines towards the marginal cost of production as more competitors enter. This is certainly the case in traditional generic drug markets where prices tend to fall from around 90% of pre-generic brand price with only one entrant to 10% or less as the molecule becomes fully commoditized. Given the need to master complicated manufacturing, to conduct some level of R&D, and to carve out a profitable market share in the absence of automatic substitution, it seems unlikely that any biologic drug will attract the number of entrants necessary to create a commodity-pricing situation. Even if this were the case, the pricing floor established by the cost of marginal production will be higher with biosimilars than with traditional generics.
The WSJ also reports that “pharmaceutical companies have been raising prices on biotech drugs about to lose patent protection to squeeze out more revenue before competition arrives…And makers of the knockoffs are setting their prices just below those marked-up ones.” A run-up in brand price at the time loss of exclusivity (LOE) is also a familiar tactic among small molecule drug makers. In fact, the phenomenon of brand drugs raising prices in the period before and after LOE is so common that economists have given it a name, the “generic paradox.” Ultimately, it is competition from other generics that will drive cost down (or lack of such competition that will keep prices high). As long as the competition is between a biosimilar and the reference drug, neither party will undermine the combined value of the market by starting a price war.
Again, none of this should come as a surprise. The relatively modest discounts associated with biosimilars as compared to their small molecule predecessors have been anticipated for some time (and demonstrated in Europe where these products have been on the market for approximately ten years). As we wrote in our 2009 report, Biosimilars: Problems, Prospects and Projections, “biogeneric producers will need to recoup R&D investments that far exceed what is required for traditional generics – and, as a result of limited competition, they are likely to have some degree of pricing power that allows them to do so…Litigation risk will also need to be considered and insured against by all parties…As a result, biosimilars are unlikely to achieve the types of societal savings provided by ANDA generics.”
A year ago, we looked at data from IMS Health showing how specialty drugs were boosting the fortunes of an industry still reeling from the body-blows of high-dollar patent expirations in preceding years. The figures were encouraging from a top-level view, but the reliance on a handful of products with questionable growth potential (Hep C treatments) left us asking if this was sustainable.
That question is still on our minds. Of course, no one thought last year’s results were a complete aberration, and the new report from IMS shows the trend continues in a healthy fashion (free download with registration). Total US drug spending grew 12.2% to $424.8 billion. The increase in the specialty drug sector was nearly twice that – 21.5% -- and, at $150.8 billion, the category now accounts for more than a third of all domestic drug spending and three-fourths of new brand drug spending. Given that this category has gone from 24% of the market to 36% in just three years, and account for 70% of spending growth over that period, IMS is right, if a bit understated, to call it “one of the most dynamic segments.”
Like last year, Hepatitis C and oncology lead the way. Along with autoimmune treatments, these products account for $19.3 billion of the overall $46.2 billion rise seen by the industry as a whole – with $6.6 billion of the increase coming from Hep C alone. By comparison, spending on oncology treatments as a whole increased by $6 billion, led by new offerings in the form of PD-1 and protein kinase inhibitors.
There are signs that the Hep C wave is cresting, however. As the authors of the IMS report state that “new therapy starts peaked in March 2015 and have started to slow as the majority of patients most in need have sought and received treatment.”
For now, we stand by our statement from a year ago, when we cautioned that the top-line figures did not necessarily herald a new Golden Age for pharma. As we stated then, “these phenomenal revenue figures are really derived from a handful of extremely successful drugs rather than a broad-based industry turnaround.”
For the first time since the FDA’s Office of Generic Drugs (OGD) was authorized to address its approval backlog with industry-paid user fees, the agency has provided a top-level view of activity in the form of its first annual report. The highlight for the generic industry is the 580 approvals (along with 146 tentative) that were issued in 2015, which includes a record-setting 99 combined approvals in December alone.
The agency, which is in negotiations with Congress to reauthorize the user fee project, can also point to dramatic operational improvements. With roughly $300 million per year provided through user fees, the agency has been able to clear almost 90% of the backlog in generic applications. The backlog, which is defined as applications that were received before October 1, 2012, initially included 2,866 ANDAs and around two-thirds as many applications for substantial changes on previously approved drugs. At the end of the reporting period, 84% of the former and 88% of the latter had been resolved in one way or another.
As is frequently the case, however, it pays to read the fine print: specifically, the most common form of ANDA-resolution by a substantial margin was “Complete Response with an Inspection.” According to the agency, this constitutes “a written FDA communication to an applicant usually describing all of the deficiencies that the agency has identified in an application that must be satisfactorily addressed before it can be approved.” If a high percentage of the sponsors of these applications take steps necessary to address the deficiencies identified in their complete response letters, a lot of the “resolved” applications could wind up back in the agency’s workload. Given the speed with which the generic industry changes, it is possible that many sponsors walked away from these applications years ago but the complete response approach does not have the same level of finality as one might expect.
Less ambiguous and of perhaps greater importance for the future is the fact that the ANDA filing backlog has been eliminated for all practical purposes. Filing can be thought of as a “pre-review review.” In other words, this is the period during which the FDA performs an initial audit of a submitted application in order to determine whether it can move forward to the substantive review phase. As recently as August 2014, more than 1,100 applications were stuck in this particularly frustrating form of limbo. In a very short period of time, the OGD addressed this backlog and is now making filing decisions in an average of 40 days, with only one percent taking longer than two months.
We came across an interesting study while doing research for a recent client project. Although it was published in 2013, we thought it was worth sharing as it has some very interesting findings on a subject that gets a lot of attention but not a whole lot of quality research: i.e. the value of me-too drugs in the US health care system.
According to the study, “in certain circumstances, me-too drugs are cost saving for payers. Namely, price discounts have to be sufficient and me-too drugs have to be launched within a few years after the first entrant. If me-too drugs are launched late, they could save payers money in the short and medium term, but could represent a cost in the long term, as they prevent conversion to low-priced alternatives after the first entrant becomes generic.” The article contains a much more detailed quantitative model of these trade-offs that may be of interest for the econometrically minded.
In addition to providing an empirical analysis of the value of me-too products from payer and societal perspectives, the study provides further evidence for a key pharmaceutical industry dynamic that we’ve noted many times over the years: products that are not first to market, a group that includes all me-too drugs by definition, have to spend disproportionate amounts on marketing just to stay in the game.
Specifically, the study reports that “on average, me-too drugs launch 2.5 years after the first entrant, with 20% more promotional investment, and capture 38% of market share within 4 years. Peak market share is significantly affected by share of voice (p<0.001) but not price discount (p=0.77).” Put differently, investing in marketing is more effective at driving market share than offering steep discounts. Whether this attempt to, in effect, buy market share through higher promotional spending, translates to a positive ROI for the pharmaceutical company is something that would vary considerably in different circumstances. (This study could, however, be used as a framework to perform the necessary calculations.)
In addition, “launch delay was significant in terms of reducing both market share (p< .001) and price (p<0.05).” Thus, while price discounts won’t drive share, the drug-maker may have no choice but to offer low prices, particularly for products that are launched well after the originator.
Another notable finding: “The time trend coefficient was negative, meaning that the first in class’s advantage increased over time. In other words, and with everything else being equal, it was better on average for a me-too drug to launch in the 1990s than in the 2000s.” The author, Stephane Régnier, hypothesizes that this may be attributable to more effective cost containment measures implemented by managed care organizations over time.
The study abstract, along with a link the full version, may be found at http://www.ncbi.nlm.nih.gov/pubmed/23440390.