Over at Healthy Algorithms, the healthy theoretical computer scientist, puts up a very interesting graph about spatial variation in Medicare expenditures in the last six months of life. Paper here. Interesting stuff, but am I the only one who gaffs when reading: “Previous studies have shown that regions with greater overall EOL spending do not have better outcomes” (EOL = end of life)?
O brave new quantum world!
Hi David and Jon,
It’s been a confusing time learning to speak like these health folks do, but I’m starting to figure it out… I think they mean that other health outcomes (like staph infections/1000 patients or something) are no better in the hospitals that spend more on end of life care.
I think that they focus on EOL spending because of the potential errors of looking at less focused spending. For example, a hospital that treats really ill patients might be spending more and getting worse outcomes than a hospital that treats not-so-ill patients, but maybe they should be spending more, since they have tougher cases to deal with, and maybe they can’t expect the same quality of outcomes because some of the people coming in a so sick that there isn’t much to be done.
EOL care partially addresses this, because maybe the patients are coming in sicker in some places than in others, but 6 months later they are just as dead.
How can you have better outcomes if you study the last six months of life? what is fascinating to me is the correlation of education and geographic heritage with end of life spending.
After working at a teaching hospital for the last 6 years my feeling is that secondary events like nosocomial Staph infections or whatever occur at a higher rate _because_ the individual is hospitalized. Did I miss something in their paper where they discussed average length of hospital stay across the cost quintiles? My guess is that higher costs are proportional to longer hospital stays which are themselves proportional to negative secondary outcomes. Cascading interventions drive up costs (much like end of life care of a pet). So there is the connection with the attitudes toward palliative care, comfort measures and hospice utilization. Clearly the message that hospitals are not god does not reach the population characterized by having the lowest %age with HS diplomas and the major conclusion from this study is that informational materials for both patients and surrogate decision makers need to be better targeted to this population.
Would also be interesting to know percentage of patients who had advanced life directives (or whatever those things are called these days) across the quintiles.