A quiet war has been slowly brewing within the U.S. healthcare system over the rising costs of new drugs. Outsiders got a rare peak into this ongoing tug-of-war between pharma companies and payers last year, when Express Scripts' (NASDAQ:ESRX) Chief Medical Officer Steve Miller helped to a broker a deal with AbbVie (NYSE:ABBV) to dramatically slash the price of new hepatitis C medications. 

Now he's looking to do the same thing with costly cancer drugs, and that could be a major blow to drugmakers such as Celgene Corp. (NASDAQ:CELG) Here's why. 

Miller's proposal would pummel Celgene's growth strategy
Miller wants to adopt a system in which cancer drugs would be reimbursed based on their effectiveness in a given tumor type. 

The idea is simple enough. Many cancer drugs are approved for multiple indications but have varying degrees of effectiveness across their approved uses. Even so, drug companies are charging the same amount for every indication, which would change, perhaps markedly so, under Miller's proposal. 

A pay-by-performance reimbursement system would thus be a huge blow to a company such as Celgene, which has thrived by continually expanding the labels of its top cancer products such as Revlimid. 

Revlimid's sales, for instance, have been growing by close to 20% annually for years, because the company has been extremely successful at expanding the drug's label for new indications such as newly diagnosed multiple myeloma:

Source: Celgene.

The drug's extensive clinical data across multiple forms of cancer, though, shows why a pay-by-performance system would be such a nightmare for Celgene and its shareholders.

Source: Celgene.

As a treatment for relapsed mantle cell lymphoma, Revlimid combined with rituximab reported overall response rates of 57% and complete response rates of 36%.  

When the drug is used in conjunction with dexamethasone as a treatment for relapsed multiple myeloma, the overall response rate comes in at close to 61%, but complete responses are only 16%, according to data from two late-stage studies. 

Digging into these data, we can immediately see that Revlimid is basically not effective whatsoever for around 40% of patients across these two particular indications. That brings up the possibility that Celgene might receive only a nominal fee, to cover manufacturing costs, in cases where the drug failed to improve clinical outcomes. 

And then there is the issue of how "performance" would be judged under such a reimbursement scheme. Revlimid's complete response rate in relapsed multiple myeloma, for instance, isn't great. Would the drug therefore only garner its full price if a complete response is observed, or would a partial response justify full reimbursement? 

Those are extremely complicated issues to work through, but they hint at multiple ways in which Revlimid's sales could crater under a pay-by-performance system. 

What's next?
Although Miller and many others are advocating for major changes in the pricing of cancer drugs as a whole, the clinical and logistical realities may preclude any dramatic modifications for the time being.

Most cancers remain incurable, meaning that patients often receive every available treatment at some point along the way. So Celgene's shareholders probably shouldn't lose too much sleep over Miller's opening salvo against pricey cancer drugs.

The cold, harsh reality is that cancer drug prices are unlikely to head lower until real competition is introduced in the marketplace. Indeed, it was the competition between AbbVie and Gilead Sciences (NASDAQ:GILD) that broke the dam on hep C drug prices. 

Unfortunately, competition is an extremely uncommon feature of the oncology landscape at present, because science is only now getting a handle on the complexities inherent in the tumor ecosystem. The good news is that there is a tremendous amount of work being done to explore the use of potent combo therapies that could one day lead to functional cures for some tumor types.

Until then, however, cancer drug prices will more than likely remain on the high side.