In the course of researching a column on end-of-life care, I came across a fascinating study attempting to test whether advance directives — the documents that spell out the sort of care someone wants at the end of their life — make much of a difference in the actual care patients receive at the end of their lives.

Michael Yanul, 58, one of the last seven long-term patients at Oak Forest Hospital in Oak Forest, Ill., lays in the bed where he has lived for the past 17 years with muscular dystrophy. (M. Spencer Green/Associated Press)

The results were surprising: The advance directives only saved money in areas where Medicare’s end-of-life costs are high. Think Florida and New York. In areas where Medicare’s costs are already low — say, Minnesota — there was no change in spending. This seems like very good evidence for the Dartmouth Atlas view of the world.

Moreover, the difference in costs was not due to a difference in treatments meant to save the patient’s life. It was due to a difference in treatments meant to prolong it.

The big difference, in fact, was that patients with end-of-life directives in high-costs areas were more likely to die outside of the hospital than patients who didn’t have end-of-life directives. That suggests the end-of-life directives do not prevent emergency measures at the end of life. If a patient is about to die, doctors try and save them. But in a non-emergency context, when doctors have time to talk with the patient’s family about treatment options, they do prevent certain types of life-prolonging treatment.

“Advance directives were originally designed to allow patients to determine in advance what treatment they would receive in the event of their incapacity,” write Doug White and Robert Arnold in an editorial accompanying the study. “Instead, it may be that treatment-limiting advance directives work not by making the decisions but by giving surrogates and physicians psychological permission to cease life support at some point.”

More on the study here. And more smart thoughts about end-of-life care — and the difficulty of knowing when the end of life actually is — here. And for more on the subject, there were lots of good comments beneath yesterday’s post on how doctors die.