Effect of post-discharge venous thromboembolism on hospital quality comparisons following hip and knee arthroplasty

J Bone Joint Surg Am. 2014 Sep 3;96(17):1476-84. doi: 10.2106/JBJS.M.01248.


Background: Symptomatic pre-discharge venous thromboembolism (VTE) rates after total or partial hip or knee arthroplasty have been proposed as patient safety indicators. However, assessing only pre-discharge VTE rates may be suboptimal for quality measurement as the duration of stay is relatively short and the VTE risk extends beyond the inpatient setting.

Methods: Patients who underwent total or partial hip or knee arthroplasty were identified in the 2008 through 2010 American College of Surgeons (ACS) National Surgical Quality Improvement Program (NSQIP) database. Outcomes of interest were the deep venous thrombosis (DVT), pulmonary embolism (PE), and overall VTE rates within thirty days after surgery and the rates during the pre-discharge and post-discharge portions of this time period. Risk-adjusted hospital rankings based on only pre-discharge (inpatient) events were compared with those based on both pre-discharge and post-discharge events within thirty days of surgery.

Results: A total of 23,924 patients underwent total or partial hip arthroplasty (8499) or knee arthroplasty (15,425) at ninety-five hospitals. For hip arthroplasty, the VTE rate was 0.9%, with 57.9% of the events occurring after discharge. For knee arthroplasty, the VTE rate was 1.9%, with 38.3% of the events occurring after discharge. The median time of VTE occurrence was eleven days postoperatively for hip arthroplasty and three days for knee arthroplasty. The median duration of stay was three days for both hip and knee arthroplasty. When hospitals were ranked according to VTE rates, hospital outlier status designations changed when post-discharge events were included (κ = 0.386; 44% false-positive rate for low outliers). The median change in hospital quality ranking was 7 (interquartile range, 2 to 17), with a rank correlation of r = 0.82.

Conclusions: Nearly twice as many VTE complications were captured if both pre-discharge and post-discharge events were considered, and inclusion of post-discharge events changed hospital quality rankings. These data suggest that inclusion of post-discharge events should be considered when comparing the quality of hospitals on the basis of postoperative VTE rates.

Level of evidence: Therapeutic Level IV. See Instructions for Authors for a complete description of levels of evidence.

Publication types

  • Comparative Study
  • Evaluation Study
  • Multicenter Study

MeSH terms

  • Aged
  • Arthroplasty, Replacement, Hip / adverse effects*
  • Arthroplasty, Replacement, Hip / methods
  • Arthroplasty, Replacement, Knee / adverse effects*
  • Arthroplasty, Replacement, Knee / methods
  • Cohort Studies
  • Confidence Intervals
  • Female
  • Humans
  • Length of Stay
  • Male
  • Middle Aged
  • Multivariate Analysis
  • Odds Ratio
  • Patient Discharge*
  • Postoperative Complications / epidemiology
  • Postoperative Complications / physiopathology
  • Quality Assurance, Health Care*
  • Retrospective Studies
  • Surgery Department, Hospital / standards*
  • United States
  • Venous Thromboembolism / epidemiology
  • Venous Thromboembolism / etiology*