End-of-Life Spending Patterns

Especially for the elderly, medical expenses often peak near the end of their lives.  Health Affairs has a study in which the authors used Medicare data from people who died in 2012 to identify four patterns in that end-of-life health spending.   (HA Article)   For Medicare, care in the last year of a beneficiary’s life represents 30% of all spending and in that last year the program spends an average $40-50,000 per beneficiary compared to an average of $7000 per year for non-decedents.     The study population was about 100,000 persons in the fee-for-service program.  (As Medicare Advantage becomes 30% or more of all Medicare recipients, it seems important to always include those people in any study of Medicare to give a full picture and identify differences between the arms.)  The researchers looked at daily health spending, medical conditions, and health services used, all for the last year of these decedents’ lives.  For the beneficiaries who died in 2012 the average age was 83, 56% were women, 88% were Caucasian, 69% did not die in a facility and 43% were in a hospice program.

Four patterns of spending were found:  high persistent; moderate persistent; progressive and late rise.  The high persistent group represented about half of the beneficiaries who died.  This group started the last year of life with high spending and it steadily increased up to death.  The moderate persistent group had moderate initial spending at the start, then a dip, then a rise in the last four months of life.  Progressives had relatively low starting spending with a continual relatively steep increase and were 10% of the study population.  Late risers, 12% of the death-bound, had very low spending until four months before death and then spending rose very rapidly.  There were only small differences by sex, age or race and ethnicity.  High persistents averaged $59,000 of health spending in the last year, while progressives averaged $37,000, moderate persistents $25,000 and late risers $11,166.  High persistents obviously also used more of just about all types of health services, especially visits to specialists, inpatient days and skilled nursing facility days.  Progressive spenders were most likely to use hospice.  High persistents also pretty obviously had more health conditions, but specific diseases were not well-associated with a particular spending trajectory.  There was geographic variation in the percent of beneficiaries falling into each trajectory.  Since it does represent such a large portion of overall Medicare spending, more research is warranted on patterns of end-of-life care, as are initiatives to lower that spending, as it pretty much is fruitless, since death ensues shortly, despite high spending.