Why I'm Not Worried About Drug Prices How I Stopped Worrying And Learned To Love The $100,000 Drug

Note: Neither Memorial Sloan Kettering nor the tool DrugAbacus are proposing any particular prices for cancer drugs. 

How much would you pay to live an extra four months with prostate cancer? This isn’t an easy question to answer, but it’s highly relevant. A prostate cancer drug from 2010 offers a nice case study.

Dendreon, the company that developed Provenge – a treatment for castration-resistant prostate cancer (survival rates are abysmal compared to overall prostate cancer) – thought that $77,000/month (according to DrugAbacus, a tool developed by Dr. Peter Bach at Memorial Sloan Kettering Cancer Center) was the right price. The UK’s National Institute for Health and Care Excellence (NICE) disagreed, and using default inputs in the DrugAbacus tool, the drug would appear to be overpriced by about $50,000 per month.

Is Provenge overpriced? The pivotal trial for Provenge found that average survival time was about 4 months longer for Provenge relative to placebo. Assuming a full-treatment cost of $93,000 (not including associated costs of physician visits), the cost per life-year gained is a fairly significant $282,000 ($93,000/0.33=$281,818). And the quality of life in those 4 months isn’t likely to be great, which means that we shouldn’t value each extra month of life as we do a month in full health. (Typically, economists use a threshold of up to $150,000 per quality-adjusted life-year [QALY] to establish cost-efficiency. Some economists argue for higher or lower thresholds.)

Does this mean that Provenge isn’t worth the money? Not necessarily. It’s very possible (and indeed, likely) that there are patients who – for reasons not entirely clear – had exceptional responses to the drug. One post-hoc analysis published in the Journal of Hematology and Oncology identified a patient with a non-clinical response (PSA level reduction) to Provenge typically seen in only 3 percent of patients in the original study. It’s not clear whether overall survival was better than the median observed in the trial.

So while the “static” analysis might reject the drug at traditional cost-effectiveness thresholds, for this patient, the cost (though still very high, objectively speaking) may very-well have been worth it. And indeed, it’s possible that Provenge will pave the way for future novel discoveries that build on the incremental benefits that it provides.  My colleagues Paul Howard and Peter Huber have written about the National Cancer Institute’s exceptional responders initiative, which is attempting to track, and understand these rare responses and develop a better understanding of when and why they happen.  If successful, biomarker-guided treatments would allow positive cost-benefit ratios for more patients.

So approaches like the DrugAbacus, or NICE, leaves us with two questions. The first is about the price of a drug, whether (given a set of assumptions about efficacy, safety etc.) it is indicative of value, and whether it is accessible to patients who need it.  For hard to treat diseases like mCRPC, innovations will likely be more incremental than breakthrough at first, which make discussions of value more complex (importantly, these incremental innovations are extremely valuable).  The technology learning curve and the cost curve look pretty steep.

But the second question, modifies the first: Are the patients receiving the drug those who are most likely to benefit as a result. If a drug regiment can be tailored so that those most likely to benefit are those who are treated, cost becomes much less of a concern.

Broadly, we can separate drugs into two segments – those intended for a large population, and those that aren’t.

Gilead Sciences Sovaldi (and the follow-on, Harvoni) are intended for the large, roughly 3 million-strong hepatitis C population in the U.S. (or more appropriately, the 170 million population of chronically infected HCV patients globally). While there are genotype variations in HCV, and one drug may work for a particular genotype and not others, the majority (if not all) American patients can be treated with some combination of existing drugs.

Provenge, on the other hand, was designed for a much smaller subset of prostate cancer patients with castration-resistant cancer. Similarly, Gleevec (which raised price concerns among many oncologists) is used to treat an even smaller patient population with chronic myeloid leukemia (where new annual cases number in the thousands).

One important reason for the uproar over Sovaldi and Harvoni’s price was that it appeared to violate an unspoken agreement between payers and drugmakers – that drugs targeting large (non-orphan) populations would be priced more moderately. (Though, one analysis of the two drugs appear to be priced within cost-effectiveness thresholds, implying that the value is worth the cost.)

Though Gleevec had its fair share of detractors, the loudest opposition came from oncologists rather than insurers and pharmacy benefit managers (PBMs). This is likely because insurers were more willing to tolerate the risk of paying for a very expensive treatment for a very rare cancer that was highly effective than they were paying for a very expensive treatment for a much more common infection.

This illustrates how the two questions play out in practice.

The primary concern with Gilead’s drugs was not only the price, but the total costs of treatment for the health care system (this became a particularly acute concern for state Medicaid agencies that have to manage with tighter budgets). Besides restricting use for drug users (those that are likely to be re-infected), targeting the drugs to the most responsive patient groups wasn’t really possible. By all accounts, the drugs were highly effective for most HCV patients.

When it came to Gleevec, however, the fact that the drug was highly targeted meant that there was less variation in value – in the original non-randomized trial, 95 percent of trial participants were alive after 18 months (before, 15-20 percent of CML patients would die each year). The right patients were receiving the right drugs, and based on default assumptions in DrugAbacus, the price for Gleevec appears to be too low.

The new science of drug development makes it easier to better target drugs to the patients who would benefit the most. But in a world where we’re moving towards value-based pricing for health care, this is also likely to result in expensive medicines.

This is where FDA reform comes into play. Of course, companies aren’t going to cut drug prices simply because time-to-market and development costs have fallen. The profit-maximizing price is the profit-maximizing price. Competition, however, can and does reduce prices within a given indication. This appears to be happening with PCSK9 inhibitors (where some PBMs are waiting for a second product on the market before adding the drugs to their formularies), for instance, and has already happened in the HCV market thanks to PBMs taking a harder stance on expensive drugs.

As FDA regulations catch up to the science, more companies will be able to get more drugs approved for any given indication. The natural result is more competition, forcing companies to accept lower prices for their drugs.

So when we talk about the “high” price of a drug, first off, it’s vital to consider the value to patients. But importantly, recognizing that the solution to high prices doesn’t have to be national price controls – rather, leaving PBMs to do what they do best, and ensuring competition in the drug space can do the same (if not better).

This piece originally appeared in Forbes.com