The least-sexy news in health care these days — bundled payments — might also be the most important.
A few years ago, under the authority of the Affordable Care Act, the federal government started an experiment: It would pay hospitals a single amount for surgeries to replace joints like knees and hips instead of paying them for each individual service. The hope was that hospitals could lower costs while maintaining or even improving the quality of care.
The program was called “bundled payments,” and it was voluntary under the Obama administration — hospitals could opt in or out. Almost immediately, it seemed to be working. Costs per procedure were going down, and quality seemed to stabilize or even improve. Hospitals didn’t have to worry about squeezing every dollar they could out of Medicare and could start focusing more on the care itself.
But some health economists had two outstanding fears: Would hospitals start performing more surgeries, which would erase any savings to Medicare? And would they start becoming more selective about their patients to avoid particularly costly ones?
According to new research led by Ezekiel Emanuel and Amol Navathe at the University of Pennsylvania published Tuesday in JAMA, the answer to both is: no, not really. The volume of procedures isn’t going up, and for the most part, the kind of patients receiving replacements isn’t changing, either.
Bundled payments got started in earnest under Obama, and now the Trump administration is embracing them too. If hospitals can lower costs while still improving the quality of care under them, then bundled payments could be a potent tool in bending the health care cost curve.
Why this new study on bundled payments is so noteworthy
Even for the researchers who analyzed the program, the findings were a bit of a surprise. You might expect hospitals to look to maximize their profits if Medicare starts paying less money per procedure. But the lack of a response they found is good news if you believe in bundled payments as one way of bending the cost curve.
The Penn researchers studied Medicare claims data in about 300 markets, split between those with a hospital participating in the bundled payment program and those without one. Under the program, hospitals are paid for a patient’s hospital stay and for 90 days after their joint replacement surgery. Providers can receive a bonus if they keep their costs down, or they could have to pay a penalty for a high-cost episode, based on a target price set by the feds.
The paper suggests that after several years, some of the biggest fears about the policy — like an increase in the total surgeries performed, which could erase any savings Medicare was seeing on a per-procedure basis — haven’t materialized. No market responses of that kind.
“My initial hypothesis was we’d see a positive response. I did not expect to find a lack of a response,” Navathe told me. “I think that’s incredibly reassuring for the policy.”
In brief, they found a negligible change in the number of joint replacements performed in markets where hospitals were participating in the bundled payment program. There wasn’t the sudden increase in surgeries that was feared.
The hope for bundled payments is that because providers receive a flat payment for a 90-day period, they will have strong incentives to perform the initial procedure in the most cost-effective way while also maintaining their quality. The idea is to avoid complications that require the patients to be readmitted, which will cost the hospital more money even though they won’t be paid more by Medicare.
Hospitals using bundled payments did notably accumulate a bigger market share, the research found, even if the total number of procedures across the market didn’t increase.
Navathe and Emanuel offered one theory to explain that. Physicians and surgeons often operate at multiple hospitals, and they may be directing more of their patients to the ones participating in bundled payments, because if they succeed in performing the surgeries at a lower cost, “you make more money on it,” Emanuel said. “Decrease complications and you’ve made your money.”
The researchers also didn’t pick up any shift in the racial, socioeconomic, or other demographic features of the patients receiving joint replacements with bundled payments, the other main fear for the policy — with one notable exception.
Patients who had previously been treated in a nursing home were less likely to receive joint replacements at hospitals participating in the program versus hospitals that were not.
“There could be some selection bias to the least sick patients and it could be partially explained by quality improvement,” Navathe said, adding that it was hard to disentangle the two.
But the prevailing finding was that hospitals receiving bundled payments did not seem to be discriminating (for lack of a better word) against racial minorities or poorer patients.
Why bundled payments could be promising for controlling other health care costs
The Trump administration is expanding bundled payments, including for knee and joint replacements. The government, in pilots started under Obama, has also begun to test bundled payments for certain kinds of cancer care and for heart stents. The next frontier is more delicate medical conditions, like congestive heart failure or pregnancy.
Obamacare opened the door for a lot of experiments in containing health care costs. Emanuel, who helped craft the law at the White House’s Office of Management and Budget, is a big believer in bundled payments. He told me his original hope was for the feds to be paying for five different procedures by bundle by 2015, though that proved too ambitious for federal officials, who worried that hospitals and doctors wouldn’t be ready.
So instead, the law created a new office within the Centers for Medicare and Medicaid Services to test some of these innovative ideas.
The evidence suggests that bundled payments have been perhaps the most successful cost experiment run under the ACA, in comparison to accountable care organizations (which pay doctors and hospitals collectively for all of a patient’s treatment and for which current evidence is more mixed) and other proposals.
Bundled payments, Emanuel said, encourage hospitals and doctors to find more efficient ways of performing these procedures. Because they receive a flat amount, any savings they achieve go right into their pocket. But bundled payments don’t require providers to “boil the ocean” in the way ACOs might, because the latter requires a much broader overhaul to find savings in treating the full range of a patient’s medical needs.
In other words, it’s easy for hospitals to figure out how to do one procedure more cost-effectively. Figuring out how to treat an entire patient more cheaply is much harder.
“I think one way of thinking about this is: After a bunch of years, ACOs save maybe hundreds of dollars per patient. Bundles, you can see savings in thousands of dollars in a year or two,” Emanuel told me. “It focuses attention in a very productive way, and you do see results relatively quickly.”
In a world of Medicaid work requirements on one side of the ideological spectrum and Medicare-for-all on the other, bundled payments aren’t the sexiest subject in health care these days.
But they received bipartisan buy-in — started under Obama and expanded under Trump — and, as this research suggests, they just might be working to help get costs under control without the unintended consequences some experts feared.