In this issue of JAMA, Cram and colleagues1 report detailed analysis of primary and revision total knee replacement (TKR) performed on more than 3.2 million Medicare beneficiaries from 1991 to 2010. This analysis confirms and quantifies several important findings. First, the use of primary and revision TKRs in Medicare beneficiaries has increased both in absolute volume from 93 230 procedures in 1991 to 226 177 procedures in 2010 and in per capita usage. Second, the pattern of care has changed substantially. Between 1991 and 2010, length of hospital stay for primary TKR decreased from 7.9 to 3.5 days and revision TKR from 8.9 to 5.0 days. However, 30-day all-cause readmission rates after primary TKR increased from 4.2% in the years 1991-1994 to 5.0% in the years 2007-2010 and after revision TKR increased from 6.1% in the years 1991-1994 to 8.9% in the years 2007-2010. A clear change in the pattern of discharge status also occurred with a shift initially to a more institutionalized setting, and recently to a discharge to home with home care services (ie, discharges to patients' homes decreased from 67.5% in the years 1991-1994 to 39.9% in the years 1999-2001 then increased to 56.2% in years 2007-2010). In addition to the increase in readmissions, the data demonstrated an important increase in infection rates for revision cases, increasing from 1.4% in the years 1991-1994 to 3.0% in the years 2007-2010.