Predicting successful outpatient total shoulder arthroplasty

Total shoulder arthroplasty (TSA) is increasingly performed safely and efficiently as an outpatient procedure in certain patients. Patient selection is often based on surgeon choice, surgeon expertise, or institutional guidelines. One orthopedic research group released a publicly available shoulder arthroplasty outpatient appropriateness risk calculator that considers patient demographic characteristics and comorbidities with the aim of helping surgeons to predict successful outpatient TSA. This study aimed to retrospectively assess the utility of this risk calculator at our institution.

Records were obtained for patients undergoing procedure code 23472 at our institution between January 1, 2018, and March 31, 2021. Patients undergoing anatomic TSA in the hospital setting were included. Records were reviewed for demographic characteristics, comorbidities, American Society of Anesthesiologists classification, and surgery duration. These data were entered into the risk calculator to calculate the likelihood of discharge by postoperative day 1. Charlson Comorbidity Index, complications, reoperations, and readmissions were also collected from patient records. Statistical analyses assessed the model's fit with our patient cohort and compared outcome measures between inpatient and outpatient groups.

Of the 792 patients whose records were initially obtained, 289 met the inclusion criteria of anatomic TSA performed in the hospital setting. Of these patients, 7 were excluded because of missing data, leaving 282 patients: 166 (58.9%) in the inpatient group and 116 (41.1%) in the outpatient group. We found no significant differences in mean age (66.4 years in inpatient group vs. 65.1 years in outpatient group, P = .28), Charlson Comorbidity Index (3.48 vs. 3.06, P = .080), or American Society of Anesthesiologists class (2.58 vs. 2.66, P = .19). Surgery time was longer in the inpatient group than the outpatient group (85 minutes vs. 77 minutes, P = .001). Overall complication rates were low (4.2% in inpatient group vs. 2.6% in outpatient group, P = .07). Readmissions and reoperations did not differ between groups. There was no difference in the average percentage likelihood of same-day discharge (55.4% in inpatient group vs. 52.4% in outpatient group, P = .24), and a receiver operating characteristic curve to assess fit with the risk calculator demonstrated an area under the curve of 0.55.

The shoulder arthroplasty risk calculator performed similarly to chance when retrospectively predicting discharge within 1 day after TSA in our patients. Complications, readmissions, and reoperations were not higher after outpatient procedures. Risk calculators for determining whether a patient should be admitted after TSA should be used cautiously because they may not provide measurable benefit over the use of surgeon experience and expertise in discharge decision making, and other factors may be relevant in the decision to perform outpatient TSA.