Establishment and validation of an interactive web-based calculator for predicting postoperative functional recovery in metatarsal fracture patients: A LASSO regression model approach

Abstract

Background Metatarsal fractures rank among the ten most common fractures.Comprehensive studies on postoperative functional recovery remain limited. A reliable predictive model for recovery outcomes is essential for optimizing patient care. Objective To develop and validate a predictive model for postoperative functional recovery in metatarsal fracture patients and implement it as an interactive web-based calculator. Methods This retrospective study included 555 metatarsal fracture patients (2018–2022), with 425 in the training cohort and 130 in the validation cohort. The outcome variable was postoperative recovery as assessed by the AOFAS midfoot scoring system. LASSO regression identified significant predictors of recovery,the selected variables underwent binary logistic regression analysis to identify independent risk factors. A prediction model was constructed using the training cohort and visualized through a nomogram. Model validation was performed internally through bootstrapping and externally using the validation cohort. The model was implemented as an interactive web calculator using R Shiny. Results At final follow-up, 71.71% of patients achieved good recovery (AOFAS score >80). The model identified ten independent risk factors, including residence location, smoking status, obesity, rehabilitation training, educational level, age, injury mechanism, infection, and anemia. The model demonstrated robust discrimination (c-index: 0.832 training, 0.838 validation) and calibration (H-L test: P = 0.994 training, P = 0.648 validation). DCA showed optimal clinical utility within 0.72–1.00 threshold probability. Protective factors included hilly areas (OR = 0.183), smoking (OR = 0.4), obesity (OR = 0.270), and undergoing rehabilitation training (OR = 0.237),while risk factors included low educational level (OR = 3.884), advanced age (OR = 2.751), high-energy injury (OR = 3.003), residence in mountainous regions (OR = 4.671), infection (OR = 16.946), and anemia (OR = 5.787). The interactive web calculator is accessible at https://metarecoverypredictor.shinyapps.io/DynNomapp/. Conclusions The validated prediction model effectively identifies risk factors for postoperative recovery in metatarsal fractures. This tool can aid clinicians in developing personalized treatment strategies and improving patient outcomes. The web-based calculator provides easy access for clinical application.