Cap Constraint

To avoid bankruptcy, hospice managers have an incentive to choose patients who are unlikely to live very long.

At the end of their life, patients may turn to hospices to alleviate the pain and discomfort associated with their illness. The cost of providing palliative care in the United States is subsidized by the social insurance program Medicare according to its reimbursement policies. This hospice benefit, which was set up in 1983, is widely regarded as a success: It improves the quality of life of patients who decide to forgo further treatment, while saving Medicare thousands of dollars that would have gone to more expensive curative care.

But the spate of hospice bankruptcies in recent years has exposed critical weaknesses in Medicare's reimbursement policies. Many hospices have been spending much more than the yearly cap on revenues set by Medicare. Overspending by several hundred thousand dollars can be devastating to a small hospice provider. The reason for this imbalance is that patients have been staying at hospices far longer than expected. In fact, the proportion of cancer patients—who typically stay a short time—has been falling, while the proportion of longer-staying non-cancer patients has been rising, and they also have been staying longer.

To keep a hospice afloat, its managers might be tempted to actively recruit and discharge certain types of patients, which goes against Medicare's objectives. "It's an undesirable practice to seek a particular patient type," says Chicago Booth professor Rodney P. Parker.

Indeed, a recent study titled "On Hospice Operations under Medicare Reimbursement Policies" by Parker with Bariş Ata of Northwestern University, Bradley L. Killaly of Emory University, and Tava Lennon Olsen of the University of Auckland, finds strong analytical evidence that the current rules for reimbursing hospices produce distortions. In particular, hospice managers tend to seek cancer and other similar patients who are not expected to live too long and discharge those who are still alive. "It was interesting to us how a fairly simple reimbursement policy was having these perverse incentives among practitioners," says Parker.

The study's results underscore the importance of reviewing Medicare's hospice reimbursement policies to prevent these unintended consequences. One potential remedy, according to the authors, is to allow smaller hospices to merge in order to increase patient volume and get a better mix of shorter- and longer-living patients.

Another solution is to give hospices an extra year to carry over any remaining cap allowance from the previous Medicare year, which runs from November 1 to October 31, so they can have a better chance of recovering their expenses. This simple modification permits a hospice to continue consuming the remaining cap that would otherwise disappear on November 1.

How Hospices Can Go Bankrupt

To be admitted to a hospice, a patient needs the signature of two physicians on a document saying that he or she is not expected to live longer than six months and that the patient agrees to forgo any curative care. Medicare pays hospices a maximum daily rate for every patient under care, which is the first part of the hospice reimbursement policy. The rates differ according to the type of care provided, but do not depend on the disease of the patient. For instance, a hospice typically receives a smaller subsidy for taking care of patients in their homes rather than in a hospice facility, but does not get more for patients with dementia compared to those with cancer. About 95 percent of hospice days reimbursed by Medicare are for routine home care.

The second part of the reimbursement policy is a payment cap applied to the entire hospice. The cap starts at zero at the beginning of the Medicare year and goes up by a fixed amount for every new patient admitted (around $23,874 per patient in 2010), and then goes down whenever the hospice receives a payment from Medicare according to the daily rate (around $147 per patient-day for routine home care in 2011). It is possible for hospices to exceed their cap because—although Medicare will keep paying them for every day that patients stay in their care—the cap does not increase if no new patients are admitted. The difference between the reimbursed amount and the hospice's cap must be paid back to Medicare. This negative cap could be sufficiently large to force a hospice into bankruptcy.

In general, hospices can go bankrupt because of several factors, the most important of which are patient volume and mix. A hospice with insufficient volume cannot cover its fixed costs and its profits will be more variable. Moreover, a hospice with too many patients who stay a long time or one with too many patients who do not stay long enough, are both at risk of going bankrupt—the former because a hospice's revenues are limited by the cap, but its daily costs of care continue; and the latter because patients die too quickly to accrue sufficient revenues to cover the fixed costs. It would be difficult for a hospice to earn more than its fixed costs if it only had patients dying from cancer, but it would also have difficulty staying under its cap if it had to care only for patients suffering from dementia.

An ideal mix of patients allows the hospice manager to leverage the benefits of both types: cancer patients who increase the cap and non-cancer patients who will allow hospices to maximize the revenues accrued under the cap. Thus, one way to ensure profitability, the authors propose, is to allow hospices to merge. A merged hospice will not only produce a larger volume of patients but, if well chosen, the right match will result in a better mix of patients. For instance, hospices in Alabama and Mississippi tend to exceed their caps because patients stay exceedingly long, while hospices in South Dakota have more patients that stay for remarkably shorter periods. Moreover, because the majority of hospice patients receive home care, there are few physical facilities that need to be managed.

Although combining hospices comes with a few complications, such as how to coordinate their shared resources, the authors feel that this is a more socially equitable solution than if hospice managers, in managing the cap, have to adjust their mix of patients by actively recruiting the "right type" or discharging patients who are still alive.

The Unintended Consequences of Managing the Cap

To avert bankruptcy, hospices with a poor patient mix may resort to actively recruiting patients who are not likely to live very long, and discharging those who have stayed too long. Indeed, the authors show that hospice managers have an incentive to engage in such tactics.

They find that as hospices become more constrained by the cap, the rate of recruiting patients who are not expected to live long—such as cancer patients—begins to dominate the recruitment of non-cancer patients, especially at the end of the Medicare year. Recruiting more cancer patients is the optimal strategy for hospice managers because these patients can bump up the cap, but are less costly to care for than non-cancer patients because they have a shorter life. Similarly, the study finds that managers trying to keep a hospice from bankruptcy have an incentive to discharge patients who are still alive, particularly non-cancer patients, and will do so more consistently throughout the year as hospices become more cap constrained.

To overcome part of the distortions created by the current rules, the authors propose allowing a hospice to continue receiving revenues for patients who are alive at the end of a year until any remaining cap from that year is exhausted. A simple way to implement this policy is to give a hospice an extra "legacy" year, which would allow the hospice to use the cap allocated in the previous Medicare year provided that the patients moving on to the next year are still alive and under its care.

While this type of policy will still generate different recruitment and discharge rates for patients with different diseases, it will at least remove managers' incentives to recruit and discharge at varying rates throughout the year. For instance, non-cancer patients will no longer be less desirable at the end compared to the beginning of the year because their contribution to the cap is carried over to the following year anyway. This proposed policy is easy to implement and can significantly increase access to palliative care for all those who need it at the end of their life.