**4. Conclusions**

Over the decades since the first deployment of ECMO its application has exponentially increased, and there has been growing interest in developing tools to guide clinical decision making, to help aid in patient selection and prognostication. While several scores have been developed, they share similar limitations secondary to the granularity of the available data, score development approaches, the focus on mortality as the main outcome of interest and reliance on data only from patients supported on ECMO. This has led to continued efforts to refine and a requirement to continuously update these scores. Future directions include a transition from a mortality focused approach to an approach focused on identifying objective short and long-term neurologic outcomes. Additionally, there is a need to develop tools capable of matching the studied ECMO cohort to a non-ECMO cohort of similar severity of illness and then develop tools capable of aiding in both patient selection and determination of the optimal ECMO initiation time to improve both mortality and neurologic outcomes.
