The most interesting article I came upon this past week dealing with an issue of great importance to older people wasn’t in JAMA or the New England Journal of Medicine and it wasn’t a report from the Institute of Medicine or from the Henry J. Kaiser Foundation. It was in the Wall Street Journal.

The article reported that beginning April 1, Medicare is embarking on a brave new experiment: it is “bundling” payment for patients getting a knee or hip replaced. MedPAC, the independent, agency that advises Congress on how to improve Medicare, has long advocated reforming the way Medicare pays for surgical procedures. And the CMS Innovation Center has funded a variety of projects testing the ability of bundling payments to improve care. But now, for the first time, proposals and theories affecting nearly half a million patients are being put into practice.

Actually, it’s not half a million patients right away. Only hospitals in the 67 metropolitan areas randomly selected by CMS will be affected—New York and Los Angeles won the lottery—hospitals that perform about one-third of all hip and knee replacement surgeries in Medicare enrollees. And calling the new payment mechanism “bundling” isn’t entirely accurate either: Medicare isn’t giving out a single lump sum for all aspects of care and telling orthopedists, hospitals, radiologists, and rehab facilities to divide it up however they see fit. What it’s doing instead is to pay everyone the way they usually do—hospitals get a single DRG (diagnosis-related group) payment, SNFs get paid a prospectively determined amount each day the patient is in the SNF, and physicians are paid on a fee-for-service basis. But if the total amount that Medicare ends up distributing over a 90-day period exceeds a target figure, the hospital has to pay back the excess. And if the total amount is less than the target, the hospital gets the difference. In short, rather than truly sharing the risk—or, from a clinical perspective, the responsibility—for care, the burden of ensuring that everyone provides optimal care rests solely on the hospital.

Now I think it’s a good idea for hospitals, rehabs, and doctors to work together—and for that matter, physical therapists and free-standing labs and radiology units as well—but I’m not convinced that placing the responsibility exclusively at the hospital’s doorstep is wise. It’s essentially the same approach taken by Medicare to the problem of hospital readmissions—of patients being discharged, only to come back to the same hospital, sometimes for the same problem, in less than a month. Medicare has instituted a system of penalties to hospitals whose readmission rates exceed a given threshold. As a result, the majority of hospitals were penalized for their readmission rates in 2015, some losing as much as 3 percent of their Medicare reimbursement. In a number of states, including New York and Massachusetts, three-quarters or more of the hospitals were hit with penalties.

The problem in both cases, the readmissions and payment for joint replacement surgery, is twofold: hospitals do not have control over all aspects of the patients care, and sometimes things go wrong that couldn’t have been prevented, no matter how much control the hospital exercised. Many Medicare enrollees are very old and very frail—these patients are likely to get sick again even if they are discharged from the hospital with follow-up arranged and their medications reviewed and a nurse visit scheduled the day after they get home, all the ingredients of a good “transitional care plan.” These same patients are likely to benefit from a stay in a skilled nursing facility or a rehabilitation hospital after they’ve had a joint replaced, strategies that cost more than sending them home with a few visits by a physical therapist and a nurse or a printout of exercises to do at home.

In the case of the new bundled payments for orthopedic procedures, the hospitals might respond by making sure that their patients only go to the very best skilled nursing facilities where they manage to restore them to perfect functioning in a matter of days or else go directly home, where they get the very best Visiting Nurse service that supplies the very best physical therapist who likewise can restore them to perfect functioning after just a few visits. But I worry that the hospitals might try to cherry pick patients—only accepting for surgery those people who are eighty-going-on-sixty and will do just fine at home with no services at all. I worry that hospitals will despair of their ability to control anything that goes on in a nursing home or home health agency and will opt instead to buy them up, leading to further consolidation within the hospital industry—and bigger isn’t always better for patients. And I worry that in the unlikely event that the system works, that care improves and costs go down, hospitals will have simply robbed Peter to pay Paul: they will achieve improvements in the domain of hip and knee surgery at the expense of care in the arena of abdominal surgery or stroke care.

I do think that older patients benefit from coordinated care. They win if their orthopedists at the hospital talk to the attending physician at the skilled nursing facility. They win if the details of their hospital stay are available electronically to the staff at the rehab facility. They win if hospitals, SNFs, and home care agencies work together. Let’s hope that Medicare’s experiment achieves that result.