We used Medicare administrative data (2002–2009) and an instrumental variables design that exploits the natural experiment created by the implementation of Medicare Part D to estimate the effect of prescription drug coverage insurance on the use and costs of inpatient services. We find that gaining prescription drug insurance through Part D caused approximately a 4 percent decrease in hospital admission rate, a 2–5 percent decrease in Medicare inpatient payments per person, and a 10–15 percent decrease in inpatient charges. Among specific types of admissions, gaining insurance was associated with significant decreases in admissions for CHF and COPD.