Predictors of Failing Same-Day Discharge after Shoulder Arthroplasty: Developing a Model to Improve Outcomes and Reduce Healthcare Cost.

Background: With increasing efforts to transition shoulder arthroplasty to the ambulatory surgery setting there is increased interest in predictive factors of failure of same day discharge (SDD). The purpose of this study was to identify predictors of failing SDD, defined as requiring at least an overnight hospital stay after shoulder arthroplasty, and to develop a predictive model to identify which patients may require postoperative hospital admission.

Methods: A retrospective review of a consecutive series of patients with rotator cuff arthropathy or osteoarthritis treated with primary anatomic (aTSA) or reverse total shoulder arthroplasty (rTSA) between January 2019 and June 2023 was conducted. Inclusion criteria included patients intended for SDD, while arthroplasty performed for fractures, patients under 45 years of age, and patients with incomplete data were excluded. Demographics, Charlson Comorbidity index (CCI), preoperative opioid use, and preoperative steroid injections were collected. Additionally, intraoperative metrics including American Society of Anesthesiologists (ASA) score, surgical/anesthesia time, surgical start time, and blood loss were collected. Multivariate logistic regression was used to identify predictors of failure of SDD. Results were displayed as odds ratios (OR) and 95% confidence intervals (CI). The alpha threshold was set to p < 0.05.

Results: A total of 333 patients (69 aTSA, 264 rTSA) were included with 92 failures of SDD (27.6%). According to multivariate results, we found older age (OR: 1.44, p=0.004), anesthesia start time per hour later in the day (OR: 6.03, p=0.017), preoperative opioid use within the past year (OR: 1.82, p=0.046), and female sex (OR: 2.76, p=0.001) as statistically significant risk factors for not achieving SDD. Additionally, each half hour increase in length of time under anesthesia was statistically significant, increasing odds of failing SSD by 4.28 per half hour (p<0.001). A diagnosis of rotator cuff arthropathy had 3.40 greater odds of failing to achieve SDD as compared to a diagnosis of osteoarthritis (p<0.001). Overall, the model had a high level of predictability, yielding a receiver operating characteristic curve area under the curve (ROC AUC) of 83%.

Conclusions: Older age, rotator cuff arthropathy, female sex, preoperative opioid use, and prolonged anesthesia exposure were significant predictors of failure of SDD after shoulder arthroplasty. These results support the utility in risk stratification strategies of patients to accurately triage them into SDD pathways. Additionally, our results suggested that collaboration between surgeons and anesthesia teams may allow for more accurate risk stratification and optimization of resource utilization when deciding which patients are appropriate candidates for SDD.