**7. Challenge and future directions**

Percutaneous cannulation for ECMO is increasingly performed by intensivists, cardiologists and emergency physicians. Despite infrequent occurrences, complications of cannulation may be associated with significant morbidity and mortality [17]. Existing simulation models mainly focus on ECMO circuit management, with less emphasis on percutaneous cannulation, and there are yet publications related to percutaneous cannulation model. Collaborative international endeavours targeted to improve cannulation safety are in progress. We have developed a simulation model for fluoroscopic-assisted dual-lumen cannulation (https://www.youtube. com/watch?v=dr02RAMRk1A) and two-cannula cannulation in veno-venous ECMO.

Despite the increasing use of simulation in many ECMO education programmes, the role of simulation remains unclear in currently available guidelines and international recommendations. The latest Extracorporeal Life Support Organization (ELSO) guideline on ECMO specialist training published in 2010 only included didactic lectures, water drills, animal laboratory sessions and bedside training as the main training modalities. Moreover, it only targeted 'ECMO specialists' (i.e. nurses, respiratory therapists, perfusionists and medical professions providing care under the guidance of ECMO physicians) and not ECMO physicians.

In a United States survey published in 2015, simulation has been adopted as part of the institutional ECMO credentialing programme for ECMO physicians in 73% of the centres, highlighting its increasing importance [18]. The EuroELSO guidelines for training and continuing education of ECMO physicians published in 2017 have incorporated high-fidelity simulation training as an alternative training modality to supplement bedside clinical hours. More importantly, its significance in teamwork and communication skills training has been acknowledged [19]. It is expected that the role of simulation will be further expanded in future versions of the ELSO guideline.

As mentioned earlier, ECMO training curriculum and thus simulation programmes are institutional specific and may be modified according to the characteristics and training needs of the institution. As a result, significant variances exist among different institutions, imposing difficulties in standardization and validation of the assessment tools. Currently, there are no validated assessment tools to assess the learning efficacy of common and essential ECMO scenarios. Joint efforts are ongoing among international ECMO educators to develop a standardized curriculum. Only through the implementation of a uniform ECMO simulation programme can multi-centred studies be carried out to provide a better understanding of the best approach in ECMO education.
