The cost of caring for America’s nearly 6 million Alzheimer’s disease patients is already $600 billion a year, factoring in the cost of uncompensated caregiving. Now, the Food and Drug Administration has approved a drug treatment that may or may not work but is set to cost $56,000 a year for the average patient – a charge that in most instances will fall to Medicare.

Biogen, the maker of aducanumab, reportedly now has until 2030 to demonstrate that the drug improves patients’ well-being. Dominick Reuter/AFP/Getty Images via TNS

Medicare should not simply cover this treatment indiscriminately. Instead, it should evaluate whether paying for it stands to help its beneficiaries.

Medicare has a research center, the Center for Medicare and Medicaid Innovation, which has the authority to determine whether paying for aducanumab, the new drug, would improve outcomes for Medicare’s millions of Alzheimer’s patients – and thus justify the cost for taxpayers and for beneficiaries, who pay premiums and out-of-pocket costs. (Under Medicare’s rules, about $11,000 of the $56,000 cost will be charged to some patients.)

At this point, nobody knows whether aducanumab actually helps patients. Usually, the FDA makes that judgment, but this drug was approved based not on a finding that it works, but on a supposition that it might. Normally, too, Medicare could rely on private insurers to help negotiate the price of the drug, but because most Alzheimer’s patients are 65 or older and covered by Medicare, the private market will have little to no influence on the price.

This is why Medicare needs to run a study, even though it will slow access to the treatment. Here’s how it would work: Medicare would cover the drug throughout the country, but randomly pick some counties where it would reimburse the drugmaker’s current price, and others where it would reimburse $0. The net effect would be that the treatment would primarily go to patients who live in the full reimbursement counties, and rarely to those who live in the $0 reimbursement counties. (Families in $0 counties may complain about not having immediate access to the drug, but there’s no reason to assume they would be worse off. This is precisely why Medicare needs to do the evaluation.)

Medicare would follow Alzheimer’s patients in both places, measuring their disease trajectories and tallying the costs of the services they often need, to determine whether access to aducanumab generally slows Alzheimer’s progression in a population. And if it does, whether that translates to financial and other benefits that could offset the drug’s up-front expense.

Evaluating whether payments are worthwhile is exactly the type of thing Medicare’s research center was created to do. The work would complement upcoming clinical studies that are supposed to learn directly whether aducanumab relieves Alzheimer’s disease.

The evidence to date shows only that it reduces the buildup of amyloid in the brains of patients with Alzheimer’s. Many previous drugs that target amyloid have failed to slow disease progression, and aducanumab’s clinical trial did not convince the scientists on the FDA’s advisory panel that it works better. Biogen, the drugmaker, reportedly now has until 2030 to demonstrate that it improves patients’ well-being. That’s too long for Medicare to wait to find out whether it should pay.

Even if the new clinical study is done sooner, the results that emerge from such trials are often poor predictors of how well treatments will work in routine patients. This is why many commercial insurers follow up FDA drug approvals with studies of real-world evidence, especially now that the FDA is accepting shakier data than it did just a decade ago.

Medicare must study aducanumab reimbursement now, and prepare to do the same whenever the FDA approves other costly, important drugs that lack evidence of effectiveness. Drug companies should know that if their medicines reach the market before the evidence is in, payment will not be automatic.