We will be performing site maintenance on our learning platform at learn.aaos.org on Sunday, February 5th from 12 AM to 5 AM EST. We apologize for the inconvenience.

AAOS Now

Published 4/1/2016

Wide Variation Found in Readmission Rates after TKA

Whether a total knee arthroplasty (TKA) patient is readmitted to the hospital within 30 to 90 days after the procedure may depend, to some extent, on the hospital's geographic location in the United States. Data from a study presented at the 2016 AAOS Annual Meeting found a wide geographic variation in readmission among hospitals across the country, with patients in the Western states having a lower 30- and 90-day readmission (RA) risk than patients in other regions of the country. In addition, hospital factors, such as hospital procedure volume, appear to play a significant role in RA rates.

"The changing landscape of health outcomes research, with its increased focus on short-term outcomes as national benchmarks for quality in arthroplasty surgery, was the impetus for the study," said first author Steven M. Kurtz, PhD, of the study, which was supported by a research grant from Stryker Orthopaedics.

In light of the variability observed among hospitals nationwide, the authors emphasized the potential for using strategies that optimize clinical pathways to reduce 30- and 90-day RA.

Obtaining and analyzing data
Medicare 100 percent national hospital claims were queried to identify 952,593 elderly patients (≥65 years) who underwent primary TKA in 3,848 hospitals between 2010 and 2013 based on International Classification of Diseases, Ninth Revision (ICD-9) Clinical Modification codes. The researchers used a 1-year look-back period before the index TKA to evaluate the following patient factors:

  • comorbidities
  • hospital volume
  • surgeon volume

The authors developed multilevel logistic regression models using clustered data structures to investigate the risk of 30- and 90-day RA, incorporating hospital, clinical, and patient factors. They studied hospital geographic location (rural/urban); bed size; and hospital type (eg, profit/nonprofit, teaching/nonteaching) as hospital factors. Clinical factors assessed included length of stay (LOS), discharge status (home vs. skilled nursing facility), and perioperative transfusion.

Evaluating results
There was a wide geographic variation in 30- and 90-day RA among hospitals. At 30 days, readmission ranged from 0 percent to 22 percent (median, 4.9 percent), whereas at 90 days, RA ranged from 0 percent to 32 percent (median, 8.6 percent). Patients in the Western states had a 9 percent to 12 percent lower 30-day RA risk and 6 percent to 11 percent lower 90-day RA risk compared to other regions of the country.

Besides geography, hospital procedure volume (P < 0.0001), rural hospital location (odds ratio: 0.96, P = 0.0272), and nonprofit ownership (odds ratio: 1.10, P < 0.0001) were the only significant hospital factors for RA among those studied.

Overall, stated the researchers, clinical factors explained more of the variation in RA rates than did general hospital factors. Use of a perioperative transfusion was associated with a 13 percent greater risk (P < 0.0001); patients discharged to home had a 25 percent lower risk (P < 0.001); surgeon volume and LOS were also significant (P < 0.0001). Specifically, shorter LOS was associated with up to 29 percent reduction of risk, with 1- to 2-day and 3- to 4-day LOS having the lowest risk, as compared to 5 or more days. In addition, each additional surgeon volume of 100 procedures was associated with 6 percent less risk.

The top five most frequently reported primary reasons for 30-day RA in TKA were related to the following surgical and medical complications:

  • wound infection (6.2 percent)
  • deep infection (4.5 percent)
  • atrial fibrillation (3.9 percent)
  • cellulitis, and abscess of leg (3.0 percent)
  • pulmonary embolism (2.6 percent)

The researchers found that the top five reasons for 90-day RA were the same as for 30-day RA.

Drawing conclusions
According to the researchers, these findings support further optimization of anti-infection measures—both intraoperative and postoperative—to reduce the broad variation in hospital readmissions. In their view, pathways to increase home discharge and diminish transfusions and periprosthetic sepsis could decrease the readmission rate.

This study has two main "take-away" messages, according to Dr. Kurtz.

"First, there is tremendous variation in readmission rates for TKA across the country," he noted. "The second is that the top reasons for readmission include infection and wound complications."

Dr. Kurtz noted that counseling patients about returning home and improving postoperative wound management are just two examples of ways to potentially reduce RAs.

"The top five reasons for readmission are factors that can be positively influenced and are all related, to some extent," noted coauthor Frank Kolisek, MD. "Three of the top five are infection-related, so it may make sense going forward to pay more attention to those patients who are at a higher risk of infection, including preoperative, perioperative, and postoperative management of these patients now that we know things such as that transfusions increase RA rates, and discharge to home from the hospital decreases RA rates, for example.

"I am not surprised by the high variation in RA rates among hospitals, because within each hospital there is high variation among the orthopaedic surgeons on staff doing the same procedure," added Dr. Kolisek. "This is difficult for each hospital to manage, but it will likely have to be done in the future, as interesting times lie ahead."

Coauthors of Drs. Kurtz and Kolisek on Paper 305, "Which Hospital and Clinical Factors Drive 30-Day Readmission after Total Knee Arthroplasty?" are Edmund Lau, MS; Kevin Ong, PhD; Edward M. Adler, MD; and Michael T. Manley, PhD. The authors' disclosure information can be accessed at www.aaos.org/disclosure