Association Between Initial Prescription Size and Likelihood of Opioid Refill After Total Knee and Hip Arthroplasty.

BACKGROUND: The present study was designed to test the hypothesis that there was no association between initial opioid prescription size and the likelihood of refill after elective primary total knee (TKA) and hip arthroplasty (THA). METHODS: We retrospectively analyzed large national datasets of commercial and Medicare insurance claims to identify a weighted cohort of 120,889 primary total joint arthroplasties (76,900 TKA and 43,989 THA) comprised of opioid-naive patients aged 18 to 75 years who had surgery between January 2015 and November 2019. The primary outcome was refill of any prescription opioid medication within 30 days after discharge, and the primary predictor variable was the total amount of opioid filled in the initial discharge prescription measured in oral morphine equivalents (OMEs). Logistic regressions were used to estimate the likelihood of refill, given a particular prescription size while adjusting for multiple patient factors, including age, sex, comorbidities, and year of surgery. RESULTS: The 30-day refill rate was 59.6% following TKA and 26.1% for THA. Adjusted odds of refill decreased by 2% for every 75 OME (10 tablets of 5 mg oxycodone) increase to the initial prescription size among the THA cohort (adjusted odds ratio [OR] = 0.98; 95% CI 0.97-0.99), and decreased by 3% for the TKA cohort (aOR = 0.97; 95% CI 0.97-0.98). CONCLUSION: These nationally representative data demonstrated that larger initial opioid prescription size was associated with small but clinically insignificant decreases in 30-day refill after total joint arthroplasty. This finding should allay concerns about efforts to decrease postsurgical opioid prescribing.

Risk factors for readmission of orthopaedic surgical patients.

BACKGROUND: Reducing hospital readmissions has become a priority in the development of policies aimed at patient safety and cost reduction. Evaluating the incidence of rehospitalization of orthopaedic surgical patients could help to identify targets for more efficient perioperative care. We addressed two questions: What is the incidence of thirty-day readmission for orthopaedic patients at an academic hospital? Can any risk factors for readmission be identified among rehospitalized patients? METHODS: This is a retrospective cohort study examining 3264 orthopaedic surgical admissions during two fiscal years from the hospital's quality-improvement database. Cases of patients with unplanned readmission within thirty days were subjected to univariate and multivariate analysis to determine the odds ratio (OR) for readmission. Further descriptive analysis was performed with use of electronic medical record data from the cohort of readmitted patients. RESULTS: The estimated cumulative incidence of unplanned thirty-day readmissions was 4.2% (i.e., 138 of the 3261 patients who were eligible for the study). Multivariate analysis indicated that marital status of "widowed" significantly increased the risk of readmission (OR, 1.846; 95% confidence interval [CI], 1.070 to 3.184; p = 0.03). Race significantly increased the odds of readmission in patients identified as African-American (OR, 2.178; 95% CI, 1.077 to 4.408; p = 0.03), or American Indian or Alaskan Native race (OR, 3.550; 95% CI, 1.429 to 8.815; p = 0.006). The risk of readmission was significant at p < 0.10 (OR 1.547; 95% CI, 0.941 to 2.545; p = 0.09) for patients with Medicaid insurance. Any intensive care unit stay gave the highest OR of readmission (OR, 2.356; 95% CI, 1.361 to 4.079; p = 0.002) for all demographic groups. Mean length of hospital stay was significantly longer, 5.9 days in the unplanned readmission group compared with 3.6 days for non-readmitted patients (OR, 1.038; 95% CI, 1.014 to 1.062; p = 0.002). Chart review of readmitted patients showed that 102 readmissions (73.9%) were classified as surgical; of these, thirty-five readmission events (34.3%) were for infection at the surgical site. CONCLUSIONS: Longer length of hospital stay or admission to the intensive care unit significantly increased the likelihood of thirty-day readmission, regardless of demographics or discharge disposition. Marital status, Medicaid insurance status, and race may indicate how a patient's social and economic resources can impact his or her risk of being readmitted to the hospital. LEVEL OF EVIDENCE: Prognostic Level II. See Instructions for Authors for a complete description of levels of evidence.