Guest column: The real cost of drug development
Pundits of drug development costs use very different models in computing the true spend in developing drugs. At one end of the spectrum is the pharma model: Take all the R&D costs over a decade, and divide by the number of drug approvals in a similar time frame. This gives an industry average of over a billion dollars per drug and includes the cost of drug failures and repeated indications before a successful one is achieved.
The Tufts Center for Drug Development tracks this carefully, and their recent estimate is $2.6 billion cost per new NME drug. It seems incredible that a pharma company may knowingly spend over a billion dollars upfront on a single development plan, but by the time the drug hits the shipping dock on the way to customers, this is a fairly realistic way of expressing the cost of development for that drug.
At the other end of the spectrum is a prediction based on the actual costs to conduct a study with the minimal number of patients for an orphan indication. A recent study in Endpoints reported that the mean pivotal trial cost was $19 million for a drug approval.
Without giving away the punchline, sadly, this is somewhat akin to claiming the real cost of driving your car is the cost to fill the gas tank. Or the cost of raising kids is just the food they eat and clothes they wear. In all three cases, nothing could be further from the truth.
Biotech/venture firms also have a point of view on the cost for clinical-stage biotech drug development to FDA approval, ie, what do we actually spend to take drugs from Phase I through approval. This approach has some convenient cost savings built in: for example, for academic start-ups, much of the primary research cost is borne by NIH and other government funding, and for pharma spin-out companies much of the early work is conveniently taken care of by the pharma before the biotech company is formed. That leaves just the costs for a clinical development program from Phase I through FDA approval.
Well, hardly simple. Sofinnova, like other venture firms that specialize in clinical-stage drug development, has learned through experience what the real cost is to push drugs from Phase 1 to FDA approval. Sofinnova tracks the ‘fully loaded cost of per-patient’ for our companies, and has done so for more than a decade. This is basically the fully-loaded costs looking at what a biotech company spends to dose each patient including the fully loaded costs (GMP manufacturing, leased space, cost of employees, and other factors).
Taking this approach — and assuming you can run a biotech company as efficiently as possible — then you take the total spend divided by the actual number of patients dosed with the drug/placebo. For example, if a biotech spends $20 million over 2 years and doses 100 patients, the fully loaded cost is $200,000 per patient. This large, fully amortized cost per patient number sometimes causes consternation in the industry as the direct clinical costs to the CRO are, say, only $5.5 million, where the remaining $14.5 million was spent on everything else: basically the infrastructure needed to do drug development: strong scientists and clinicians, GMP drug supply, toxicology studies, and the electric bills that keep the lights on. Vacation pay, employee bonus and health plans, business travel, IPO and fundraising costs.
If they are doing things right, toss in the December holiday party, and journal club costs. It is these fully amortized costs that add up quickly.
We first computed the fully loaded cost per patient math circa 2005. As our own biotech portfolio was still growing, we informally solicited data from dozens of clinical companies funded by brand name venture firms, including several brand-name, Sand Hill Rd firms, and combined them to make a confidential dataset of several dozen, clinically mature companies, yielding the following compiled data:
Average company spend = $78 million
Average number of patients = 402 (geometric mean average)
Average per patient cost = $168,000.
We felt this was shockingly high. When we examined oncology companies only, the average cost per patient was even higher, $258,000, and for protein therapeutic companies it was $345,000 per patient.
Although this subset of biotech companies was limited at the time, the message was unmistakable: The cost to run a venture-backed, clinical stage biotech for a few years, dosing hundreds of patients (which is typically a very aggressive number required for FDA approval) is certainly not $19 million.
In the last decade, we have had 17 FDA drug approvals come out of Sofinnova-funded companies. Three of these companies were acquired before FDA approval, and so we don’t have full insight into the total cost of development for these companies.
Nonetheless, the remaining 14 companies that took their drugs all the way to FDA approval collectively raised/spent $4.65 billion, giving an average cost per drug to approval of $327 million (+/-264 million, SD).
Venture-backed biotech companies are fairly efficient at developing drugs, and we believe this is part of the reason why the biotech industry has boomed for more than two decades.
Bottom line: Drug development is an expensive business, but those that can do it more efficiently and cheaply than others should be able to stay in business.
So why do we do it? Why do we spend so much on developing drugs, and investing in the quality of life, for ourselves and our children? Many things in life are more expensive than they might seem on face value, including the car you drive, the children you raise, and the life-saving drugs you take. Yet all provide a quantum change in quality of life, despite the occasional flat tire, the diapers and cost of college and, yes, the cost to demonstrate drug efficacy and safety, held to one of the highest standards imaginable: FDA approval.
Mike Powell is a general partner and Jason Pitts is an associate at Sofinnova Ventures.