Outpatient Management of Operative Ankle Fractures and Its Benefits: A Predictive Model for Patient Selection.

Background Outpatient management of operative ankle fractures has consistently demonstrated financial and clinical benefits in the literature. We aim to determine factors associated with preoperative and postoperative complications of ankle fractures and create a predictive model that will assist surgeons in identifying patients optimized for outpatient operative management. Methods We performed a retrospective analysis of adult patients presenting to a single Level I trauma center from 2019 to 2024 with ankle fractures requiring open reduction and internal fixation (ORIF). The primary outcome was any complication, and secondary outcomes were preoperative complications and postoperative complications. Complications were then compared to patient characteristics using univariate analyses, followed by multivariate logistic regression to identify independent predictors of complications. Results A total of 318 patients were included, with 78 experiencing one or more complications. Time to surgery, kidney disease, diabetes, drug use, chronic obstructive pulmonary disease, cardiac comorbidities other than heart failure, and nicotine use were significantly associated with complications either prior to or following operative fixation. Time to surgery, diabetes, and drug use were associated with preoperative complications. When screening for the likelihood of preoperative complications, model performance demonstrated adequate discrimination and calibration, with an area-under-the-curve of 0.78, balanced accuracy of 0.75, sensitivity of 0.86, specificity of 0.64, positive predictive value of 0.20, negative predictive value of 0.98, and a Hosmer-Lemeshow p-value of 0.63. Conclusion Time to surgery, recreational drug use, and diabetes were found to be significantly associated with preoperative complications. The resulting model can reliably identify patients at low risk of preoperative complications with a sensitivity of 0.86and a negative predictive value of 0.98. Treating physicians may utilize this model to select patients optimized for safer outpatient management.