Around 25 percent of US Medicare spending occurs in the last year of life. Does that mean society is throwing large amounts of medical treatment at some patients in a futile, if noble, effort to extend lives that are bound to end soon?

A study coauthored by economics professor Amy Finkelstein offers a resounding answer: No.

After examining millions of medical records, the researchers found that although Medicare spending is concentrated among people who end up dying, little of it goes to patients whose death within the year was judged highly likely. For example, less than 5 percent is applied to patients in the single highest-risk percentile—and their predicted one-year mortality rate is still just 46 percent.

“Very little money is spent on people who we know with high probability are going to die in a short amount of time,” says Finkelstein.

The study used machine learning to evaluate the impact of a range of variables, producing a probability of death within one year for each patient. It revealed that fewer than 10 percent of people who die in a given year have a predicted one-year mortality rate over 50 percent. And even when people are admitted to a hospital in their last year of life, fewer than 4 percent have a predicted one-year mortality rate of 80 percent or higher at the time of admission.

The study shows that even relatively low-mortality health scenarios for the elderly will include some deaths—not that individual treatment decisions represent long-shot cases.