Measuring Returns to Hospital Care: Evidence from Ambulance Referral Patterns

We consider whether hospitals that receive higher payments from Medicare improve patient outcomes, using exogenous variation in ambulance company assignment among patients who live near one another. Using Medicare data from 2002-10 on assignment across ambulance companies and New York State data from 2000-6 on assignment across area boundaries, we find that patients who are brought to higher-cost hospitals achieve better outcomes. Our estimates imply that a one standard deviation increase in Medicare reimbursement leads to a 4 percentage point (or 10 percent) reduction in mortality; the implied cost per at least 1 year of life saved is approximately $80,000.