Predictive factors for mortality in patients admitted to the intensive care unit

Mortality prediction is a resource that allows optimized allocation of resources in the intensive care unit (ICU). Available scores provide an easy approach to mortality prediction in ICU patients. However, evidence has shown that these models are inadequate and imprecise for the intended purpose. Therefore, novel investigations are needed to optimize these models by including new variables. On the other hand, it has also been proposed to create new predictive models based on novel technology. Machine learning, artificial intelligence, models have been proposed as an alternative to solve this issue. However, more validation studies are needed to recommend further this novel approach toward mortality prediction. This review aims to analyze predictive mortality factors in the ICU context that can further optimize available scores or serve as an independent predictive factor.