Hospital Readmission and Length of Stay Over Time in Patients Undergoing Major Cardiovascular and Orthopedic Surgery: A Tale of 2 States

Abstract

Background: Readmission and length of stay (LOS) are increasingly accepted as quality measures for surgical care. Centers for Medicare & Medicaid Services will soon assess penalties for excessive readmissions after coronary artery bypass graft (CABG) surgery and hip and knee replacements.

Objective: To determine and compare population level changes in LOS and relationship with 30-day readmission over time for patients undergoing CABG and hip and knee replacements. Secondary objective was to determine relationship between LOS and discharge disposition as well as mortality.

Research design: Observational cohort study of patients undergoing CABG and hip and knee replacements in New York and California. Temporal trends in LOS, discharge disposition, 30-day readmission, and mortality were examined. Generalized linear-mixed models, accounting for hospital clustering, were used to assess differences in outcomes.

Subjects: Patients undergoing CABG and hip and knee replacements in New York and California between 2005 and 2011.

Measures: Trends in LOS, discharge disposition, 30-day readmission and mortality, and risk-adjusted odds of all-cause 30-day readmission.

Results: We identified 206,784, 336,271, and 416,391 patients who underwent CABG, hip, and knee replacements, respectively, in New York State and California between 2005 and 2011. The risks of readmission within 30 days decreased over time in both states. LOS decreased by 1 day after hip and knee surgery and remained unchanged after CABG. Adjusted analysis confirmed these trends. In secondary analyses patients in New York had higher overall odds of 30-day readmission compared with patients in California.

Conclusions: We found no evidence of inverse relationship between LOS and readmission over time. In hip and knee replacement there is strong evidence that both LOS and readmission have been reduced simultaneously.