The final year accounts for a far smaller share of total health expenditures than would-be reformers believe.
In the quest to reduce health spending without harming health outcomes, one area has always loomed large: end-of-life care. But it turns out that the share of total outlays that occurs near the end of life is small, and in any case it is difficult to predict when that stage is near, as a new analysis shows.
This issue is typically presented as the ultimate Ebenezer Scrooge dilemma: huge opportunities for reducing costs but at significant moral cost. The misleading debate over non-existent “death panels” as part of the Affordable Care Act followed this pattern. Yet both dimensions of the story are flawed because the savings from altering end-of-life care are more modest than conventional wisdom suggests, and many people actually prefer lower-cost, at-home options to dying in a hospital.
First, on the potential for cost savings, historical figures showed that more than 25 percent of total Medicare expenditures occurred in the last year of a beneficiary’s life. That suggested significant opportunity to wring out costs from the care delivered near the end of life. The share of Medicare spending on those at the end of their lives has, however, declined somewhat over time.
More importantly, as Zeke Emanuel and Linda Emanuel pointed out more than 20 years ago, about 5 percent of Medicare beneficiaries die each year, compared with about 1 percent of the total population. As a result, end-of-life care represents a smaller share of total health-care spending than of Medicare spending. Indeed, the most recent estimates suggest that less than 9 percent of total U.S. health-care spending occurs in the final year of life. And that share is actually lower than in many other developed countries.
A new study points to an even more challenging issue: We have little idea who is on the verge of death, so knowing after the fact that a certain amount of spending occurred in the last year of life doesn’t provide much insight to practitioners or policy makers. A team of economists at Stanford, the Massachusetts Institute of Technology and Harvard used machine-learning algorithms to predict which Medicare beneficiaries would die in the next year.
They find that most people who die aren’t expected to, and many people who are expected to die don’t. In particular, less than 10 percent of those who die within 12 months had a predicted mortality above 50 percent. And if beneficiaries are ranked by their predicted mortality, the group with a high chance of dying accounted for only 5 percent of total Medicare spending, and among them about half survived in any case, perhaps in part due to the health care they received. As the study’s authors conclude, “These findings suggest that a focus on end-of-life spending is not, by itself, a useful way to identify wasteful spending.”
That conclusion doesn’t obviate the second aspect of the debate over end-of-life care: what patients actually prefer, regardless of the cost implications. The evidence shows dramatic and largely unwarranted variation in end-of-life and other types of care. Surveys of patients, furthermore, suggest that the vast majority (perhaps as much as 80 percent) of us would prefer to die at home. The evidence shows, though, that we are as likely to spend our final days in an acute care hospital as at home. Expanding the use of advance directives, which specify patient preferences for their own care, could help rectify this imbalance. And some new business models, including one that has generated much attention recently, are devoted to delivering the care that people say they want as illnesses advance.
The bottom line is that the Scrooge dilemma doesn’t really exist: End-of-life spending involves less than 10 percent of total expenditures, and trying to eliminate high end-of-life spending is an impractical target in any case because death is mostly unexpected a year ahead of time. Where we can focus, though, is on allowing patients to choose their care at different stages — and that isn’t necessarily the high-cost care provided in a hospital at the very end of life.
~Peter R. Orszag is a Bloomberg Opinion columnist.