164 Chapter 9 bedside staff tend to accept a slightly higher oxygen saturation, whereas a machine will always follow its programmed instructions. Titration by a machine may lead to stricter adherence of the target range, and can therefore lead to a different median SpO2 and spread around the median. The NeOProM studies included a significant overlap in the achieved oxygen saturation distribution. This may have diluted the effect of choice in target range on clinical outcomes such as ROP, NEC and mortality. Strict titration by automated technology can aid in finally solving the puzzle of the most appropriate range to target. We are currently conducting a study comparing an SpO2 TR of 91%-95% with an SpO2 TR of 92%-96% during automated oxygen control. We expect that a set target range of 92%-96% will result in a more stable SpO2 and reduction of hypoxic episodes (SpO2 <80%) due to the position at the oxygen-haemoglobin dissociation curve. Finally, further research is warranted to elucidate the effect of AOC on clinical outcome, preferably in a very large, carefully designed randomised controlled trial with continuous automated oxygen control from the most effective device during the entire admission.