Predicting 1-Year Survival Using Machine Learning in Very Old Patients Before ICU Admission

Discussions about the benefits of admitting very old individuals to intensive care unit (ICU) remain challenging. We hypothesized that data-driven algorithms could leverage extensive real-life data to provide more accurate long-term predictions. Our objective was to evaluate the ability of machine learning (ML) algorithms to predict the one-year survival rate of very old patients prior to their admission to an ICU.