Thursday, July 21, 2016

An Update on Costs of End-of-Life Care

For those interested in the health care costs in end-of-life care, Medicare data are the obvious place to look.Of the 2.6 million people who died in the U.S. in 2014,  2.1 million, or eight out of 10, were people on Medicare, making Medicare the largest insurer of medical care provided at the end of life. Spending on Medicare beneficiaries in their last year of life accounts for about 25% of total Medicare spending on beneficiaries age 65 or older." Juliette Cubanski, Tricia Neuman, Shannon Griffin, and Anthony Damico make this point at the start of their short "data note" entitled "Medicare Spending at the End of Life: A Snapshot of Beneficiaries Who Died in 2014 and the Cost of Their Care" (July 2016, published by the Kaiser Family Foundation). "

Average health care spending for Medicare recipients who died in 2014 was $34,529, nearly four times as high as the average Medicare spending of $9,121 for those who didn't die. This general pattern isn't surprising: after all, those who die often tend to have health issues beforehand.  The detailed data shows that the biggest part of this cost difference is driven by higher spending for in-patient care in hospitals for those who died in 2014. What's interesting to me is that the share of Medicare spending going to those who die in that year seems to be diminishing.

Figure 3: The share of total traditional Medicare spending on traditional Medicare beneficiaries who died at some point in the year has declined over time

What explains this shift? The report lists these causes:
"In addition, we find that total spending on people who die in a given year accounts for a relatively small and declining share of traditional Medicare spending. This reduction is likely due to a combination of factors, including: growth in the number of traditional Medicare beneficiaries overall as the baby boom generation ages on to Medicare, which means a younger, healthier beneficiary population, on average; gains in life expectancy, which means beneficiaries are living longer and dying at older ages; lower average per capita spending on older decedents compared to younger decedents; slower growth in the rate of annual per capita spending for decedents than survivors, and a slight decline between 2000 and 2014 in the share of beneficiaries in traditional Medicare who died at some point in each year."
(A couple of notes here: 1) The graph and all the data here refer to "traditional Medicare," which is the two-thirds of Medicare recipients who are not in "Medicare Advantage" plans. In traditional Medicare, the government pays health care providers on a fee-for-service basis, and thus has good data on what the costs were for services each year. In Medicare Advantage, Medicare makes monthly insurance-like payments to a managed care organization--like a health maintenance organization--and so the government does not have readily available data on the costs of what actual health care was provided at any given time. 2) The 13.5% in the graph above for 2014 doesn't match the 25% at the top. The difference is that this figure looks at the health care costs incurred in 2014 for those who died in 2014. The 25% figure refers to health care costs incurred in the 12 months before death--which usually reaches back into the previous year. For looking at trends, either approach can work fine, but plotting data for costs in the 12 months before death and comparing it to other spending in the same time interval is a more complicated tas, and the official data is organized on an annual basis, so that's what is reported here.)  

A misconception which seems popular, at least based on the kind of questions I hear, is that end-of-life spending is especially high for the very elderly. That doesn't seem to be true. This figures shows spending of those who died in 2014 by age: for example, Medicare spending on 65 year-olds who died in 2014 averaged $38,840, while for those over age 100 it averaged $14,985. Conversely, Medicare spending on those who survived 2014 tends to rise by age.
Figure 9: Medicare per capita spending for decedents over age 65 declined with age in 2014, while spending for survivors increased
This pattern seems like a positive one to me, in the sense that I suspect there is more that health care can do for the average person who is 65 or 70, compared to the average person who is 100 or 105. A more detailed breakdown of this data shows that when just looking at health care costs of those who died in 2014 by age, those who were in their late 60s had much higher expenditures on in-patient hospital costs (the orange bars), while the older age groups tended to have higher spending on hospice or skilled-nursing facility care. 
Figure 10: Medicare spending declined with age for decedents over age 65 in 2014, mainly due to lower inpatient hospital spending

End-of-life patients do tend to be high-cost patients, and in general terms, that pattern seems appropriate. But I've written before that a main goal for end-of-life care, shared both by many patients and health care professionals, is to make greater use of hospice, skilled-nursing, and at-home care at the end of life, rather than intensive care units in a hospital setting. The evidence shows that over time, the costs of end-of-life care are a diminishing share of US health care spending, and it is consistent with the belief that a shift toward greater use of hospice and other options at the end of life is gradually underway.