Last week, the administration announced that it would not finalize the demonstration project designed to test payment changes to drugs in Medicare Part B. Part B covers outpatient drugs, those that are administered in a doctor’s office or a hospital. Payments for drugs under Part B are made directly to providers based on the sales price — without any consideration of effectiveness or any formulary management. The pilot would have changed the payment rate from 6 percent of average sales price (ASP) to 2.5 percent plus a flat fee. A second phase would have tested the use of value-based purchasing tools. The demonstration was beset by criticism from the outset: from pharmaceutical companies, specialty doctors groups, and some patient groups. The proposal even faced pushback from Democrats in Congress. After initially defending it, the administration eventually decided to drop it.

But this demonstration was just that: a demonstration, a test. The authority given to the Center for Medicare and Medicaid Innovation (CMMI) only allows the center to expand and spread policies that are shown to reduce costs without decreasing quality or that improve quality without increasing costs. These policies cannot deny or limit benefits. The Part B pilot was designed to test the results of changing payments to doctors. If even the threat of testing meaningful changes provokes this kind of response, it doesn’t bode well for efforts to reduce public sector health care costs in the future. Despite widespread, bipartisan agreement that we should be paying for value, even with respect to drugs, the uproar over Part B shows there remains strong resistance to specific policies that attempt to move towards that goal.

Demonstrations Should Test Meaningful Policy Changes

It’s troubling, but not altogether surprising, that lobbying and political pressure successfully outweighed evidence-based policy. Or in this case, even the threat of an attempt at creating evidence-based policy. Pharmaceutical companies successfully mounted a multi-million dollar campaign against this pilot. The claims of those opposed included that patients would be harmed and might not have access to life-saving drugs. Others in the medical community supported the pilot, noting that this has already been done in the private sector.

Testing new policies in health care can be difficult, especially when it comes to payment and changing incentives. Testing changes to coverage and payment in a randomized trial is the exception rather than the norm. That’s why the RAND Health Insurance Experiment and Oregon Health Experiment continue to be relevant years later. And unlike biomedical research, we can’t test for the side effects of payment reform on mice. We can model the potential results but ultimately the changes need to be tested in a real world environment.

If cutting health care costs were easy, surely we would have done it by now. Health care is now more than one-sixth of our economy and it continues to grow. The Institute of Medicine (IOM) says there’s $750 billion of waste in the system, nearly one-third of health spending. Many tests and procedures are performed unnecessarily, a problem that physicians recognize. But the solution is not straightforward. Cost-cutting measures are likely going to be unpopular because someone—doctors, hospitals, drug companies—is going to make less money as a result.

Take the example of the “Cadillac” tax, another policy left in the cost savings graveyard. A 40 percent excise tax on plans costing more than $10,200 for an individual and $27,500 for a family, it is projected to save $87 billion over 10 years and over time to incentivize employers to offer less expensive plans. But other than economists, the tax has very few supporters. Its implementation has been delayed two years until 2020 and it seems unlikely to survive, given the political climate.

Current Demonstrations Haven’t Made Significant Payment Changes (Yet)

CMMI has piloted a number of demonstrations on bundled payments, hospital payments, and accountable care, and MACRA moves Medicare physician payment towards a value structure. But we aren’t there yet. So far, most demonstrations have not made significant changes to health care spending. Participants have had little downside risk while receiving additional bonuses for meeting certain quality and savings targets. While this approach is key to drawing in participants at the outset, it shows little will among providers to risk profits.

While CMS has reported the ACO program generated savings in spending on beneficiaries, on net the program still costs CMS money because those savings are offset by shared savings payments. As of the end of 2015, only three ACOs in the Shared Savings Program were participating in Track 2, the track that involves downside risk. Programs with more risk like the Pioneer ACO program have shown some savings from the organizations that survive, but in the Pioneer case, all but nine of the original 32 participants dropped out after they lost money.

Finally, it bears repeating that paying for value means that some things inevitably offer low relative value. Not every intervention and every drug can be high value. (Don Berwick underscored this point when he noted that only in the fictional Lake Woebegone could all children be above average.) Determining what is high value and what is low value is certainly the tricky part and that is where the conflict will likely arise. What determines if something is high value? There are a number of possibilities, including patient preferences, various quality measures, clinical effectiveness, etc. — and some of those may conflict. An entire other post could be (and has been) written about the definition of value. But so far we don’t have a definition that everyone in the health system agrees on.

The future of CMMI is uncertain in a new administration. Lawmakers on Capitol Hill have come out strongly against mandatory demonstrations, one of the tools central in the center’s ability to spread any innovation it finds that saves money or improves quality. They argue that the demonstrations exceed the center’s authority and that we can move towards value-based payment without them. We may soon test whether this theory is true, as well as whether support for value-based payment is just a lot of talk.