"Why Is End-of-Life Spending So High? Evidence from Cancer Patients"
joint with Ran Balicer, Liran Einav, Amy Finkelstein, Tzvi Shir, and Salomon Stemmer

Abstract: A large share of healthcare spending is concentrated at the end of life. While this fact partly reflects higher spending on sicker patients, who are also more likely to die, spending on decedents is higher than spending on survivors even conditional on ex-ante mortality risk. We use rich data from Israel on cancer patients to provide new evidence for why dying is expensive. First, we describe where spending is coming from. Inpatient admissions account for more than 90% of the difference in risk-adjusted spending between cancer survivors and decedents, with half of this difference due to unplanned admissions. Such admissions occur predominantly in the 1-2 last months of life, and spending on them does not depend on ex-ante mortality risk. Second, we look at spending determinants. Similarity between US and Israeli spending patterns suggests that US institutions might not be the main drivers of elevated spending on decedents. Additional evidence suggests that spending is driven by both preferences and technology.