Evaluating AI and traditional models for injury prediction and prevention in high-contact sports

This systematic review and meta-analysis, registered under PROSPERO and following PRISMA 2020 guidelines, evaluates the efficacy of artificial intelligence in predicting and preventing injuries within high-contact sports. By synthesizing data from prospective cohort and observational studies across major databases, the research compares machine learning-based models against traditional predictive methods regarding injury incidence, type, and recovery outcomes. The study utilizes a structured PICO framework to analyze model accuracy, robustness, and interpretability, aiming to bridge the methodological fragmentation in sports medicine. Ultimately, the qualitative and quantitative findings provide insights into individualized risk factor identification and the practical utility of AI for enhancing athlete safety.