Data has shown 25% of Medicare spending occurs during the last year of beneficiaries' lives, leading some to speculate that Medicare is wasting money on end-of-life care—but a new study published in the journal Science is challenging that assumption, Orly Nadel Farber writes for STAT News.
For the study, researchers used machine learning to examine the relationship between mortality among Medicare beneficiaries and Medicare spending. The researchers found that beneficiaries who had the highest one-year mortality risk accounted for less than 5% of overall Medicare spending—much lower than the previous estimate of 25%. What's more, the researchers found that beneficiaries with the highest mortality risks were more likely than not to survive for longer than one year. In essence, "predicting life and death has odds similar to flipping heads or tails," Farber writes.
According to Farber, the researchers were so surprised by their findings, they wondered if there was a problem with the algorithm they had used. To check their findings, the researchers created a so-called "oracle," which is an algorithm that weighed both real and predicted deaths. The researchers found that, even with the new algorithm, their results remained relatively the same. According to the researchers, beneficiaries with the highest morality risks accounted for a relatively small amount of overall Medicare spending, while about 30% to 50% of total spending rates were largely associated with how severely ill a beneficiary was.
Ultimately, Amy Finkelstein, an author on the study and an economist at the Massachusetts Institute of Technology, said the findings show that "[w]e spend money on sick people," and "some of them die," while "some of them recover." She added, "Maybe some recover, in part, because of what we spent on them."
The findings, Farber writes, challenge the idea that "there is a good way to distinguish the sick and dying patient from the sick patient who will survive."
Similarly, Stephanie Harman, who works in palliative care and biomedical ethics at Stanford University, said the study highlights a major issue with using data on known deaths to examine spending on end-of-life care. The issue being that researchers often say spending was wasteful in retrospect. "Just because someone is seriously ill with an uncertain prognosis doesn't mean that their health care spending is wasteful," she said.
Harman added that the "quality of the care that's delivered and the patients' quality of life" should be considered when evaluating spending on end-of-life care.
Finkelstein seemingly agreed, saying researchers should look through each intervention individually to determine which ones produce health benefits for patients and which do not. Finkelstein added that, while there is plenty of wasteful spending in health care, eliminating care in the final year of a patient's life might not be the best way to resolve the problem. "We need to be a little more careful and a little more sensible when leaping from fact to conclusion," she said (Farber, STAT News, 6/28).
Advisory Board's take
Deirdre Saulet, Practice Manager, Oncology Roundtable
This study adds to the ongoing dialogue around end-of-life care in America, of which spending is just one aspect. As the study's authors underscore, death is unpredictable. The medical system should work with patients to ensure that their needs will be met in light of unpredictable outcomes.
A study in Health Affairs found that only one-third of Americans complete any type of advanced directive. Notably, this rate didn't increase significantly when the authors looked at patients with chronic illness—in fact, only about 40 percent of cancer patients had completed an advance directive. To truly improve end-of-life care, providers need to prioritize these challenging conversations, make sure patients understand their prognosis, work with them to develop goals for their care, and respect those choices.
The University of Rochester Medical Center (URMC) did this through an intervention in which social workers met with cancer patients to guide them through a list of questions about their goals for treatment, expectations for care and preferences for end-of-life decisions. Social workers helped patients identify their top priorities and coached them on how to bring up these issues with their care team. In a randomized control trial, URMC showed that patients who had received the coaching were twice as likely to ask end-of-life-related questions with their care team at their next visit.
We've collected the questions and conversation prompts that URMC used. Make sure you view this patient question prompt list for end-of-life care to start improving end-of-life care for patients.