177 10 Summary In the future research on this topic will involve a direct head-to-head comparison of all ventilators, preferably under the same conditions. This could be achieved by testing all available algorithms against a model of a preterm infant. This will allow us to inform clinicians what to expect when using an algorithm/ventilator combination. Directly following my PhD, I started this project at the University of Tasmania in Australia under supervision of Prof Peter Dargaville with a team of clinicians and engineers. Furthermore, automated oxygen control gives us the tool to elucidate what the least harmful range to target is. Previous studies have struggled to provide this evidence as during significant overlap between the compared target ranges occurred, reducing discriminative power. Automated oxygen control will provide strict titration and prevent this overlap. We are currently testing an SpO2 target range of 91%-95% against one of 92%-96%. Finally, further research should be done to provide strong evidence on the effect of automated oxygen control on clinical outcome, preferably using the most effective automated oxygen controller.