The purpose of this study was to analyze the hospital, clinical, and patient factors associated with inpatient readmission after total knee arthroplasty (TKA) in the Medicare population and to understand the primary reasons for readmission.
The Medicare 100% national hospital claims database was used to identify 952,593 older patients (65+) with a primary TKA in 3848 hospitals between 2010 and 2013. A multilevel logistic regression analysis with a clustered data structure was used to investigate the risk of all-cause 30- and 90-day readmission, incorporating hospital, clinical, and patient factors.
At 30 days, readmission ranged from 0% to 22% (median, 4.9%), whereas at 90 days, readmission ranged from 0% to 32% (median, 8.6%). Geographic census region, hospital procedure volume, rural hospital location, and nonprofit ownership were the only significant hospital factors among those we studied. Evaluation of clinical factors showed use of a perioperative transfusion was associated with 13% greater risk; patients discharged to home had 25% lower risk; and surgeon volume and length of stay were also significant. These effect sizes were at least comparable to patient factors, such as age, gender, comorbidities, and socioeconomic status. The top 5 most frequently reported primary reasons for 30- or 90-day readmission in TKA were surgery and medical related: wound infection, deep infection, atrial fibrillation, cellulitis and abscess of leg, or pulmonary embolism.
The results of this study support further optimization of anti-infection measures, both intraoperative and postoperative, to reduce the broad variation in hospital readmissions.