Predictive Factors Influencing Discharge Destination Following Spinal Cord Injury

SCI is a crippling illness that can significantly lower a person's quality of life. The severity and location of the injury determine the extent of functional disability experienced by the patient. Consequently, SCI patients often require an extended hospital stay for both acute care and rehabilitation. In our research, we have proposed the use of Gradient boosting, a machine learning algorithm, to accurately predict the likelihood of specific events. This algorithm constructs a series of models, each aiming to correct the errors of the previous one. Gradient boosting has shown promise in forecasting a number of outcomes, such as the location of discharge following a spinal cord injury. To predict the likelihood of SCI patients being released to non-routine settings, like long-term acute care hospitals or skilled nursing institutions, we used a gradient boosting technique in our study. Our model incorporated several predictive factors, including injury severity, level of injury, completeness of injury, age, social support, and insurance coverage. Impressively, our model achieved a commendable level of accuracy in predicting the discharge destination.