Quick…when you hear the words “prescription drugs”, what questions pop into your head? Perhaps: What’s with the crazy price tag? How well do they work really? What’s the difference between treatments sold by Pfizer and Genentech versus those peddled on late night infomercials? Why is it taking so long to find a cure for cancer? Assuming we do find the cure someday, can we afford it? Is Cialis right for me?
As medicinal progress continues pace forward, life expectancies and demand for life-enhancing medications correspondingly increase. Alongside this progression will be a myriad of economic and ethical dilemmas that must be solved. In order to intelligently ponder these dilemmas and make rational choices, a fundamental understanding of the nature of prescription drugs and drug development is needed. The purpose of this article is to provide some of this understanding, and to bring to light some of the more interesting dilemmas.
The discussions will focus on the following questions:
- Why is it so challenging to bring a drug to market?
- Why does it take so long?
- Why does developing drugs cost so much?
- How truly effective are these expensive “breakthrough” drugs on the market?
- What are the dilemmas I call “the 3 E’s” and how are they interrelated?
The Daunting Difficulty of Bringing a Drug to Market
In 1992, I joined a biotechnology company named Athena Neurosciences which later morphed into Elan Pharmaceuticals. Athena was focused on finding therapies for Alzheimer’s disease, and at the time they were in the process of identifying the lead molecule to use for human clinical trials. Now in 2012, Elan Pharmaceuticals, along with partners Johnson & Johnson and Pfizer, are still in late-stage clinical trials for Alzheimer’s with no guarantee of ultimate success. In other words, after 20 years and counting, they’re still working on it. Why is bringing a drug to market so difficult? The answer: because God made us so danged complicated.
How many times have you seen an evening news story touting a potential breakthrough cancer/AIDS/heart disease/Alzheimer’s/etc. treatment that may be on the horizon? The report usually ends with a statement like, “Scientists caution that this has only been tested in mice so far, and it’s still many years away from potentially reaching market.” Indeed, if there was money to be made in finding effective medical treatments for rats, there would be many more amazingly successful biotechs. Unfortunately, translating animal study successes to humans has proven quite challenging for two main reasons: one, although animal models may be predictive of human effect, it is far from a perfect model. Second, there’s much greater ease and flexibility with testing in mice versus testing in humans, primarily because mice don’t have access to the same quality lawyers and lobbyists that humans do. To complicate things further, even humans are often not great models for other humans: what’s safe and effective for some people may not be safe and effective for others.
As such, there is (and rightly so) a long and arduous formal drug development process that must be followed in order to bring a drug to market. After a compound is identified and tested in animals for safety and efficacy, it must undergo smaller “Phase 1” safety trials at relatively low doses to show that it can be safely administered to humans. These early small trials also must determine how the drug interacts inside the body: how it is absorbed and distributed (pharmacokinetics) and what effect the drug has on the body (pharmacodynamics). The next step is to run some larger “Phase 2” clinical trials to determine whether the drug may actually be effective on the target disease. These will also try to determine what the maximum tolerated dose (or “MTD”) is, as well as the dose that may give the best trade-off in terms of efficacy and safety. Comedian Jerry Seinfeld joked about the “maximum dose” label on medicines, saying, “That’s right, find the dose that will kill me…then back it off a little.” Fortunately for the clinical trial volunteers, the cut-off point for MTD is the point significant side effects are observed, not point of painful death. The degree of “acceptable” side effects depends on a number of factors, including the severity of the target disease. For instance, more side effects would be tolerated if the drug was targeting terminal cancer versus one designed to lower cholesterol. If all goes well in Phase 1 and Phase 2, the company may decide to conduct large (and tremendously expensive) “Phase 3” trials to prove the drug really works. Phase 3 trials, also referred to as pivotal trials, typically consist of many hundreds or thousands of patients (the FDA website guidance suggests that for Phase 3 trials, “The number of subjects usually ranges from several hundred to about 3,000 people”). These trials must also be “placebo controlled”; that is, patients receiving study drug must actively compared to patients receiving a non-active substance or an alternative existing treatment. This is to ensure that any observed favorable outcome from study drug patients was not due to placebo effect or random chance. The company must also decide how many study patients to enroll for the Phase 3 trials: the higher the number of patients, the better the chance the studies will statistically show the drug works (assuming it works), but the more expensive the trial gets. To make patient enrollment even more difficult, there are usually strict criteria on which patients may be acceptable for a study. For instance, a breast cancer study doesn’t simply include all women with breast cancer; it may only accept patients who haven’t already been given certain other types of treatment, whose cancers are at a certain stage, whose cancer cells test positive for certain proteins, and so forth. Indeed, patient enrollment is one of the most time-consuming activities and biggest factors of prolonged drug development timelines. As a rule of thumb, it generally takes about 5-10 years for a drug candidate to move through the three trial phases, and sometimes significantly longer if hiccups occur.
What are the odds of success? Biotechnology Industry Organization (BIO) released results of a study in 2011 demonstrating that, between 2003 and 2010, only about 9% of drugs entering human clinical trials eventually achieved FDA approval. Also in 2011, the Centre for Medicines Research reported study findings showing that since 2008, drug candidates reaching Phase 2 testing had a failure rate of 82%, and those reaching Phase 3 had a failure rate of 50%. Given that different diseases have varying complexity levels, these percentages can vary significantly based on target indications. But the point is that for the most part, even very promising early-stage drug candidates face low odds of eventually getting approved. So the next time you see a late-night TV infomercial miracle cure that promises to heal your chronic pain, cure your asthma and fix your back acne, ask the following questions before shelling out the twenty bucks: (a) Did this product go through the vigorous clinical trials process required by the FDA for approval; if not, (b) may I trust the product to work based wholly on the anecdotal testimonies of ten vaguely compensated people who look like the infomercial host’s cousins?
Drug Development’s Enormously Expensive Price Tag
In 2001, the Tufts Center for the Study of Drug Development released a long-awaited and subsequently widely-cited report declaring that, based on its research, the estimated all-inclusive cost of bringing a drug to market was $802 million (in 2000 dollars). This was compared to Tufts’ previous estimate of $231 million in 2007. Tufts cited the significant rise in the cost of clinical trials as the major culprit in the increase. Although there is some debate over the $802 million number, we can at least agree that developing drugs is enormously expensive.
There are a multitude of components that make up the total world of drug development expenses, including clinical trial and manufacturing costs, research and preclinical testing, regulatory fees, and a host of others. The biggest component is usually the notoriously expensive cost of running human clinical trials. To get a feel for why clinical trials are so costly, let’s focus only on two specific items: expenses related to patient testing, and those related to site monitoring visits.
Patient testing expenses relate to the protocol of procedures that each patient must undergo in the trial. Some common activities include physical exams, blood draws, drug administration, questionnaires and informed consent forms, etc. Depending on the target disease indication, there could be various X-rays, MRI’s, CT scans, ECG’s, and other pricey procedures. Also, the sponsor must pay for investigator fees (the supervising physician), study coordinator fees (usually a nurse managing the trial for the site), pharmacy fees, and patient expense reimbursements. Finally, some larger medical centers charge an institutional overhead fee that typically ranges around 15%-30% of the aforementioned fees. These patient / investigator site costs vary according to type of study, duration, frequency of treatment cycles, and type and location of study sites. However, a typical range may be $10,000 to $30,000 per patient per study.
Another big trial expense is what’s called site monitoring costs. Each site needs to be periodically “monitored”; that is, a trained trial professional must visit each site to ensure accurate recordkeeping and proper adherence to study protocol. This is important because even if a site has enrolled patients who are perfect fits in terms of trial criteria, the data could be rendered moot and the patients disqualified from the study if the site did not strictly follow proper protocol. Not only are there usually a number of planned monitoring visits to every site over a study’s course, the number of required sites have significantly increased over the years. As an ever-increasing number of clinical trials are initiated, competition for qualified patients has grown. This has exacerbated the need to sign up greater numbers of sites to achieve target patient numbers, as well as the need to seek investigators in more countries. Monitoring visits require travel expenses as well as labor costs for not just the hours spent monitoring (usually about 4-6 hours per site), but hours spent travelling as well.
Just to demonstrate the potential scope of the aforementioned two cost categories: assume a phase 3 trial is planned consisting of 1,500 patients enrolled by 200 investigator sites. Assume expected patient testing costs of $20,000 per patient. For monitoring, assume each site will require 8 monitoring visits (including initiation and close-out visits) over the study’s duration. Also assume each visit will require 4 hours on-site monitoring time, and an average of 8 hours round-trip travel time, or 12 hours total labor hours per visit. Further assume a fully-burdened labor cost of $75 per hour for study monitors. Finally, assume travel expenses (airfare, hotels, meals, car rentals, etc.) of $1,000 per trip. Here are the totals:
- Patient testing/site expenses: $30 million (1,500 patients x $20K/pt)
- Monitoring expenses (labor): $1.44 million ($75/hour x 12 hours per trip = $900/trip; 200 sites x 8 visits per site = 1,600 monitoring visits total; $900/trip x 1,600 visits = $1.44 million)
- Monitoring expenses (travel): $1.6 million ($1,000/trip x 1,600 visits)
Just these two trial components will cost over $33 million, and these do not include a myriad of other required items such as data management, biostatistics, and project management, just to name a few. Additionally, if overseas trial sites are included, travel expenses will significantly increase due to higher travel times and airfares. Also note that clinical trials expense is only one of many cost components required for developing a drug. Finally, note that this hypothetical trial is only one of several clinical trials required for potential approval. Did I mention drug development is expensive?
So we know that, even without “price gouging” fatty cost layers, companies will still charge big bucks for drugs if they are to cover drug development costs (which include the cost of trial failures) and target a reasonable profit. We also know that, without compelling profit potential, investors won’t fund risky biotech companies with only promising treatment ideas but low odds of success based on historic industry statistics. So a key question is: can we lower development costs, which in theory would lower costs for marketed drugs? If you look at the cost components in the previous examples, the task may seem daunting: how much can we lower the cost of an ECG or MRI? Or airfare to investigator sites in Ukraine? Or the cost of service providers like CRO’s (Contract Research Organizations) to help run clinical trials? The amounts of savings that can be squeezed out of such items are likely quite limited. Perhaps the most promising potential cost-saving developments are those centering on ways to make clinical trials more efficient. As we gain better understanding of diseases over time, we may find more efficient ways to identify and target those diseases, which should lead to better trial designs, better identification of qualified study patients, and shorter study timelines. This, in turn, should lead to lower trial costs and longer profitable patent life.
For instance, there are continually ongoing efforts to identify appropriate biomarkers for complicated diseases. A biomarker is simply a characteristic or substance that indicates the presence of a condition or disease. High cholesterol, for example, is a biomarker for heart disease. My graying hairs and increasingly creaky bones are biomarkers for my advancing age. In the medical world, it’s much more difficult to find good biomarkers. Let’s use Alzheimer’s research as an example: historically, the only way to confirm that a patient has Alzheimer’s is by performing an autopsy. Since the prospect of an autopsy might discourage patients from participating in a clinical trial, it would be useful if a less invasive biomarker-based screening methodology could be used. Additionally, having an accurate biomarker screening test would not only ensure the right patients are enrolled for a study, it would also eliminate costly screen failures, i.e. patients that are initially tested and evaluated but subsequently disqualified from the trial. As it turns out, researchers have identified potential Alzheimer’s biomarkers, such as the measurement of two proteins found in a person’s cerebral spinal fluid: the amyloid beta protein and the tau protein. Research has indicated that low levels of amyloid beta and high levels of tau in cerebral spinal fluid have a strong correlation to the presence of Alzheimer’s disease. If an accurate, timely, and economical screening methodology can be developed for testing levels of amyloid beta and/or tau in a patient’s cerebral spinal fluid, it would be a major boon towards more efficient diagnosis of Alzheimer’s disease. This is just one example of researchers’ significant efforts underway to find better disease biomarkers.
Additionally, the FDA may consider more efficient and streamlined requirements for drug approvals under certain circumstances, similar to its current “accelerated approval” and “priority review” processes. For example, a proposal is currently being considered that would allow “breakthrough” therapies, defined as experimental drugs that show a very significant effect early in development for serious or life-threatening diseases, to obtain FDA approval with fewer and shorter clinical trials as well as fewer patients. The FDA has already indicated its support of the proposal, which is currently being considered by Congress. If this and similar streamlined approval processes are feasible and implemented, it could materially reduce the costs and timelines of drug development.
Modest evidence of drug efficacy
Complicating the drug cost debate is the fact the many approved drugs are shown to only have relatively modest effect on disease patients, at least on a median basis. For instance, the discussion isn’t about a drug that cures cancer; it’s about a drug that may prolong life by a few months or may prevent the cancer from progressing for a while. To be sure, a few additional months of life are a very big deal, and significant in terms of clinical trials results. But it does make the discussion more complicated, especially given that these study results are often quite fuzzy.
First, let’s review the basic statistic concept of “median”: median is middle number in a group of numbers. So for a group of numbers consisting of 1, 3, 4, 6, 7, the median is 4. But also note that the median number doesn’t reflect outliers. Suppose a study group of seven patients given a study drug showed the following overall survival benefit in months, respectively: 1, 1, 3, 90, and 150. The median benefit would be 3, but that number is arguably misleading, yes? Granted this example is simplistic and extreme, but the point is a median number may not provide a full picture and can over-state or under-state the underlying potential.
With that in mind, let’s look at some further issues relating to study result observations. In 2010, the FDA approved the novel therapeutic cancer vaccine Provenge, which in a Phase 3 trial demonstrated that prostate cancer patients given Provenge survived a median of 4.1 months longer than placebo patients. The 4.1 months was significant and resulted in the drug being approved, but subsequent analysis suggests that the survival difference may be even bigger. It turns out that two-thirds of the placebo patients were given a frozen version of Provenge called Frovenge (the protocol dictated that if the placebo patient’s cancers progressed, they would be given Frovenge). If one assumes that a portion of the placebo patients lived longer as a result of Frovenge, this could suggest that the survival gap between placebo patients and Provenge patients should be even greater if Frovenge didn’t enter the picture. A subsequent analysis of the placebo patients showed that those getting Frovenge did live longer than those that did not, although this does not provide conclusive evidence of the theory (for one thing, it’s thought that the placebo patients not given Frovenge were generally sicker, and that could be a reason for their shorter survival). But the key question is how should a patient evaluate the potential benefit of taking Provenge? Should they expect to live four months longer taking Provenge, all else equal? Or is the expectancy significantly more than four months? Could they be one of patients along the spectrum for whom Provenge has no effect? Or one who could benefit much more than the 4.1 month median survival increase observed in the study? The point is that it’s very difficult to individually quantify the drug’s potential benefit. The specified effect based on study observations is more a guidance than anything concrete.
Another example is Avastin, one of the world’s best-selling cancer drugs. When Avastin was first approved for metastatic colorectal cancer, the main basis for approval was a pivotal Phase 3 trial that showed patients on Avastin + IFL (a chemotherapy consisting of irinotecan, fluorouracil, and leucovorin) demonstrated median overall survival of 20.3 months, compared to 15.6 months in patients taking placebo and IFL. However, a later trial comparing Avastin + FOLFOX4 (a more current chemotherapy consisting of folinic acid, fluorouracil, and oxaliplatin), the respective median overall survivals were 13.0 months versus 10.8 months, respectively. Does this apparent reduction in overall survival months between the two studies (from 4.7 months to 2.2 months) make Avastin any less compelling as a colorectal cancer treatment? Does this influence discussion over whether Avastin is worth a $50,000 per year price tag? Should cost be a consideration for a possible additional day or two months or four months of life?
For those who answered that you can’t put a price tag on possible prolonged life, consider the following scenario also involving Avastin. In 2008, Avastin received conditional approval for treatment of breast cancer based on a clinical trial that showed patients on Avastin + paclitaxel had a median of 5.5 months more progression-free survival than patients on paclitaxel alone. Note that “progression-free survival” does not mean living longer; it means that the patients’ cancers did not progress as quickly while on the drug. The condition of the approval was that sponsor Genentech must conduct further clinical trials, and at least one of those trials must confirm efficacy; that is, confirm the significant effect on progression-free survival, or show a significant effect on overall survival. Unfortunately, subsequent trials failed to meet any of those goals. In fact, the previously observed median progression-free survival of 5.5 months dropped to 0.8 months. As a result, in November 2011 the FDA revoked Avastin’s approval for breast cancer. This means that although physicians at their discretion may still prescribe Avastin to breast cancer patients, insurance companies are less likely to cover the $88,000 per year cost. To some breast cancer patients and physicians, this was an alarming development. Their argument is that despite what the clinical trials did or did not prove, Avastin likely helps a subset of breast cancer patients, and to rescind the approval could mean taking away the last treatment hope for some patients. So here the key questions are: does a demonstration of progression-free survival, which entails a quality of life enhancement rather than extension of life, justify a drug that not only causes a higher rate of toxicities and serious side effects, but costs $88,000 per year? Despite failing clinical trials, should a drug for serious conditions like cancer be granted approval status if it may help a subset of patients, if only theoretically?
The Ethics, Efficacy and Economics Dilemmas (aka “The 3 E’s”)
Early in my career, I worked for Titan Pharmaceuticals, a small biotech company trying to develop vaccines for cancer. When the clinical trials failed, I watched the value of my company stock fall off a cliff and cursed my fate…until something came to light that put my selfish plight in perspective. We became aware of a St. Louis woman named Elizabeth Grubesich, who was a patient in our clinical trial for TriAb, a vaccine for metastatic breast cancer. After she began receiving doses of TriAb, her condition improved so dramatically that she went from a state of despair to swimming regularly, taking botany and history classes, and gardening. When the clinical trial failed, Titan made the decision to stop all further development for the drug. It seemed the reasonable decision: the clinical trial showed that 68% of the participating women saw their cancer get worse, and there was no clear evidence that the drug was working. But to Elizabeth Grubesich, who was convinced that TriAb was not only saving her life but allowing her to enjoy living, this was devastating news. In response, she set out on a mission to save the drug and her own life (see The Wall Street Journal article detailing her quest here). Ms. Grubesich lobbied Titan to continue development of the drug; she tried contacted patient advocacy groups, the National cancer Institute, the media, and even the original inventor of the drug, in an effort to either encourage Titan to continue development or perhaps find another company that would. When she spoke to the FDA, she admonished a public-health specialist: “Don’t tell me you’re looking out for my interests. If you’re looking out for my interests, you’d approve TriAb.” Titan had agreed to give one year’s supply of TriAb to her on a “compassionate use” basis free of charge (“compassionate use” is a program in which a company is allowed to provide unapproved drugs to patients who have no other alternatives), but when that was used up there would be no more drug available. So on one side, you have a company that, despite extensive clinical trials, had no tangible evidence the drug actually worked and prudently decided to stop development. On the other side, you have a person who is convinced that the drug works for her and is desperately trying to save her own life. Yet drug approvals must be based on demonstrated scientific evidence that’s objective and statistically validated. Individual anecdotal accounts, no matter how dramatic, are not evidence of efficacy, are factored into approvals, and may not in themselves offer a compelling reason for further development. Ms. Grubesich’s doctor, who called her response to TriAb “remarkable” and was instrumental in helping her obtain additional supplies of TriAb, said, “We admit we don’t understand why some people do well on this vaccine.”
Like Ms. Grubesich, Avastin breast cancer patients were upset and alarmed at the prospect of losing their drug as a treatment option. Although Avastin had failed to meet its clinical trials efficacy goals and had demonstrated 13% higher toxicity rate, some Avastin breast cancer patients were convinced the drug was effective for them and perhaps their last remaining effective treatment option. A number of these patients showed up to the FDA advisory committee meeting and urged the advisory panel to maintain Avastin’s approval status for the indication. However, four breast cancer advocacy groups testified in favor of withdrawing the approval, citing lack of demonstrated efficacy and increased toxicity for breast cancer patients. After Christine Brunswick, the vice president of the National Breast Cancer Coalition and herself a cancer survivor, spoke in favor of withdrawing the drug for breast cancer due to the toxicities and lack of demonstrated trial efficacy, the next speaker was applauded when she said, “I am disgusted to have to speak after that woman.” When the lone patient representative on the advisory panel, Natalie Compagni-Portis, cited that Avastin’s clinical trials failed to show an improvement in quality of life, an Avastin-taking patient in the audience displayed a photo album and exclaimed, “What about this quality of life? Five hundred photos of my travels since 2003!” After the final vote in which the panel voted unanimously 6-0 to withdraw Avastin’s approval for breast cancer, patients in the audience erupted with angry shouts towards the panel. One patient shouted, “What do you want us to take? We have nothing else!” Another yelled at the patient representative panelist, “What a patient representative! You better hope your breast cancer doesn’t come back. You’re an embarrassment to all cancer survivors.”
The conundrums are substantial. Companies cannot feasibly continue to develop drugs that show little or no evidence of efficacy; but on the other hand, even in light of failed clinical trials, it cannot be definitely determined that the drug does not work on at least a subset of patients. Patients may have experienced wonderful results while on a particular drug despite its trial failures; however, it can’t be definitively concluded it was the drug that was the true cause of the effect and not some other factor. Perhaps, given the negative trial results, the patient might be better served with an alternative treatment that might be more effective with less bad side effects, but can the patient afford to take that gamble when his/her life literally depends on it? It would be terrific if the FDA would approve drugs sooner or ease their strict approval criteria so that patients can have access to potential life-saving treatments faster or at all; but that could increase the potential for ineffective and/or unsafe drugs to get through the system, thereby putting patients at risk for toxic side effects or by preventing them from choosing alternative treatments. As potential consumers, we want to have access to the best and most effective drugs available should the need arise; but as payers and complainers of increasing medical insurance premiums, will we be willing to pay ever-higher premiums, assuming we can afford insurance at all? Finally, although the FDA is currently forbidden by law to consider drug costs in their approval decisions, can society afford to continually pay for expensive drugs coming to market that only demonstrate marginal medical benefits? Does the answer to this question change if the expensive drugs coming to market demonstrate incredibly excellent medical benefits?
I would love to have good answers to the aforementioned conundrums…to share with you as well as to hire myself out as a charismatic and high-priced consultant. Because these answers are at this time beyond my limited intellectual capacity, the purposes of this article are simply to provide a degree of fundamental understanding of the world of prescription drugs, and to elicit some food for thought. With reasonable recognition of the significant challenges associated with drug development and the “3 E” dilemmas, we can enhance our efforts to productively deal with those challenges in the future.
More from chris:
Drug Development and Pharmaceutical Companies: Why Are Getting New Medicines to Market so Painfully Difficult?