Abstract
BACKGROUND
Risk stratification tools are being increasingly utilized to guide patient selection for outpatient shoulder arthroplasty. The purpose of this study was to identify the existing calculators used to predict discharge disposition, postoperative complications, hospital readmissions, and patient candidacy for outpatient shoulder arthroplasty and to compare the specific components used to generate their prediction models.
METHODS
This review was conducted in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analysis protocol. PubMed, Cochrane Library, Scopus, and OVID Medline were searched for studies that developed calculators used to determine patient candidacy for outpatient surgery or predict discharge disposition, the risk of postoperative complications, and hospital readmissions after anatomic or reverse total shoulder arthroplasty (TSA). Reviews, case reports, letters to the editor, and studies including hemiarthroplasty cases were excluded. Data extracted included authors, year of publication, study design, patient population, sample size, input variables, comorbidities, method of validation, and intended purpose. The pros and cons of each calculator as reported by the respective authors were evaluated.
RESULTS
Eleven publications met inclusion criteria. Three tools assessed patient candidacy for outpatient TSA, 3 tools evaluated the risk of 30- or 90-day hospital readmission and postoperative complications, and 5 tools predicted discharge destination. Four calculators validated previously constructed comorbidity indices used as risk predictors after shoulder arthroplasty, including the Charlson Comorbidity Index, Elixhauser Comorbidity Index, modified Frailty Index, and the Outpatient Arthroplasty Risk Assessment, while 7 developed newcalculators. Nine studies utilized multiple logistic regression to develop their calculators, while 1 study developed their algorithm based on previous literature and 1 used univariate analysis. Five tools were built using data from a single institution, 2 using data pooled from 2 institutions, and 4 from large national databases. All studies used preoperative data points in their algorithms with one tool additionally using intraoperative data points. The number of inputs ranged from 5 to 57 items. Four calculators assessed psychological comorbidities, 3 included inputs for substance use, and 1 calculator accounted for race.
CONCLUSION
The variation in perioperative risk calculators after TSA highlights the need for standardization and external validation of the existing tools. As the use of outpatient shoulder arthroplasty increases, these calculators may become outdated or require revision. Incorporation of socioeconomic and psychological measures into these calculators should be investigated.
Collapse