There is general agreement that the United States spends too much on health care, especially on pharmaceuticals. But what we spend on drugs is not simply a function of price. If eggs double in price, people can simply cut the number of eggs they eat in half. Simply stated, cost is the product of (price per unit times the number of units purchased).
Drugs are not eggs. It would not be rational for patients to respond to price hikes by taking half of their usual medications. But what if scientific evidence showed that half the usual dosage was just as effective and had lower toxicity? A rational and well-informed patient would have every reason to cut their dose in half, thus reducing the cost by 50%, which would also reduce potential side effects.
In 2019, we coined the term “interventional pharmacoeconomics,” suggesting that the development of evidence-based off-label dosage regimens could markedly reduce spending on drugs. Proof of concept had already been demonstrated with a randomized trial of low-dose abiraterone, demonstrating that a 75% dose reduction yielded equivalent efficacy. National Comprehensive Cancer Network subsequently supported this approach.
A confirmatory trial from India was presented at the 2025 ASCO meeting. This lower-dose approach is widely used internationally, thereby increasing access in some countries.
Abiraterone is not a one-off. Multiple studies have demonstrated that an 88-94% reduction of nivolumab (i.e., 20 mg every two or three weeks) is effective. This was first demonstrated by its manufacturer, Bristol-Myers Squibb, prior to FDA approval, in metastatic renal cell cancer (RCC).
The primary objective of this 168 patient study was “To evaluate the dose response relationship in the 0.3, 2, and 10 mg/kg BMS-936558 arms as measured by PFS.” But there was no evidence of a dose-response relationship over this greater than 30-fold range! While this protocol was drafted more than a decade prior to FDA’s Project Optimus, it is exactly the type of trial recommended by FDA in its 2024 Guidance, and alone should support the use of this lower dose, at least in RCC. (There are also recently completed studies from India demonstrating that similar doses are effective: head and neck cancer and multiple solid tumors.)
Given that CMS spent nearly $2 billion on nivolumab in 2023, CMS should be motivated to encourage the use of a lower dose, perhaps simply 40 mg (the smallest vial) administered every four weeks, rather than the labeled dose of 480 mg every four weeks. It doesn’t take a health economist to figure out that if this change were mandated by limiting CMS coverage to a reasonable and necessary dose, that spending on this drug would drop by 92%, or $1.84 billion. Furthermore, a 92% reduction in nivolumab dose could also reduce the severity of potentially life-threatening immune-related adverse events, which would secondarily decrease costs attributable to treatment of such toxicities, often requiring off-label use of expensive immunosuppressants.
There is an even larger opportunity to reduce spending on pembrolizumab, on which CMS spent over $5.4 billion in 2023. Pembrolizumab is sold as 100 mg vials, and the standard dosages are 200 mg every three weeks or 400 mg every six weeks. Given the perceived clinical interchangeability of nivolumab and pembrolizumab, it is highly likely that a dose of 100 mg every six weeks would be effective, as supported by previously published simulations. This would reduce spending on pembrolizumab by 75%, a savings of over $4 billion annually.
There are also clear examples of opportunities to reduce spending on Medicare Part D drugs, with sotorasib being the best example. This drug was approved in 2021 at a dose of 960 mg daily, accompanied by a post-marketing requirement to compare the approved dose to a lower dose because of the evidence that the absorption of the drug was very poor at higher doses. That trial did indeed demonstrate that the lower dose of 240 mg daily had an improved therapeutic index. Still, the recommended dose was not modified because of an apparent concern that noninferiority of the lower dose had not been proven—even though that was not the objective of the post-marketing requirement.
Project Optimus at the FDA’s Oncology Center of Excellence is already attempting to “reform the dose optimization and dose selection paradigm in oncology drug development.” The obvious next step is to turn these findings into Medicare payment policy by applying the long-standing Medicare requirement that coverage be limited to that which is “reasonable and necessary” to drug dosing. Medicare already does this for cancer screening; it will only pay for one screening mammogram per year, and a screening colonoscopy once every 24 months for high risk beneficiaries, and once every ten years (or 48 months after a previous flexible sigmoidoscopy) for low-risk beneficiaries. CMS should do the same for drug dosing.
Framed differently, a reasonable question is whether it’s appropriate for Medicare to continue covering sotorasib at the full 960 mg daily dose when evidence suggests a 75% lower dose may be equally effective. Since sotorasib is sold in bottles that each contain a 30 day supply of the labeled dose, Medicare Part D could simply limit refills to once every four months. Coverage changes to Medicare Part D oral drugs are complicated by the lack of a national formulary. However, changes to Medicare Part B payment policy for parenteral drugs can be made through evidence-based National Coverage Determinations.
It’s important that CMS have sufficient evidence before imposing such changes.
This presents another opportunity for HHS. Randomized clinical trials of lower-dose regimens to generate additional evidence can be cost-savings, since the decreased drug cost in the lower-dose arm more than covers the cost of the trial. Even if only 20% of the patients in such a trial were CMS beneficiaries, the trial would pay for itself if the lower-dose arm reduced drug costs by 50% or more. These trials could be conducted by NCI, using its existing NCTN infrastructure. Scientific input could be provided by content experts in oncology and clinical pharmacology (including but not limited to FDA).
When justified, the results could be utilized to modify the maximum “reasonable and necessary” dose that CMS will fund.
These changes in CMS payment policy aren’t just cost-saving—they give us the opportunity to do right by patients. In the case of sotorasib, it would significantly reduce the burden of side effects like disabling diarrhea. Patients shouldn’t have to endure chronic grade 2 diarrhea (five to seven additional bowel movements a day), which is primarily due by unabsorbed drug. A lower dose offers the same clinical benefit with less toxicity, better quality of life and lower out-of-pocket drug costs. There’s every reason to believe that similar dose reductions for other drugs would have similar patient benefits.
Our current strategy is to prescribe and pay for the labeled (often the maximum tolerated) dose – when what we need to be doing is deploying optimal dosing. Post-marketing dose-ranging clinical trials will improve prescribing of select oncology drugs, thereby reducing side effects, patient co-payments, and Medicare’s financial health.
And that is how we can make America healthier.
