I feel you’re wrong on two directions. Sure, most people who die weren’t obviously on end of life. But those people tend not to be the ones with high end of life costs. The high cost is due to their circumstances, not the label. Second, the folks we are identifying as end of life tend to have a pretty good true positive : false positive rate. So if you’re in that bucket, you probably ought to be there.
My opinions don’t matter. Argue against the paper cited in the article “Predictive modeling of U.S. health care spending in late life” https://pubmed.ncbi.nlm.nih.gov/29954980/ Abstract:
That one-quarter of Medicare spending in the United States occurs in the last year of life is commonly interpreted as waste. But this interpretation presumes knowledge of who will die and when. Here we analyze how spending is distributed by predicted mortality, based on a machine-learning model of annual mortality risk built using Medicare claims. Death is highly unpredictable. Less than 5% of spending is accounted for by individuals with predicted mortality above 50%. The simple fact that we spend more on the sick-both on those who recover and those who die-accounts for 30 to 50% of the concentration of spending on the dead. Our results suggest that spending on the ex post dead does not necessarily mean that we spend on the ex ante "hopeless."
> That one-quarter of Medicare spending in the United States occurs in the last year of life is commonly interpreted as waste.
This is the issue this paragraph is dedicated to refuting. But this thread is not about this issue. So it’s largely just an irrelevant passage. However;
> Here we analyze how spending is distributed by predicted mortality, based on a machine-learning model of annual mortality risk built using Medicare claims. Death is highly unpredictable. Less than 5% of spending is accounted for by individuals with predicted mortality above 50%.
This quote is dumb. As a data scientist I give it an F. Less than 5% of spending is accounted for by individuals with predicted mortality above 50%? Ok but what proportion of individuals is that? Is it 10% in which case they’re underrepresented? 5% in which case they’re average? Or .1% in which case they’re pulling in 50x their share? If we pull the criteria down to 40% what happens? What share of people who died are above 50% in the model? It’s a quote that is basically impossible to make use of on its face but is being used to push a very specific narrative.
