AbstractAimClinical quality improvement is often cumbersome due to established protocols. We aimed to investigate whether outcomes of preterm infants improve with protocol revisions using iteration cycles.MethodsPreterm infants born <28 weeks gestation between January 2006 and December 2015 were retrospectively analysed. Protocols were revised using Plan Do Check Act cycle. Death and serious adverse events at term were reviewed in six‐monthly quality improvement meetings. Adverse outcome of death or motor/sensory impairments at two years was compared before and after two major protocol changes, which were implemented in January 2008 and January 2012.ResultsBased on the appraisal for period 2006–2007, strategies for surfactant, narcotics, parenteral nutrition, respiratory gas humidity and prophylactic indomethacin and antibiotics were changed for period 2008–2011. For period 2012–2015, stabilisation of infants was accelerated via very early catheterisation. Of 162 infants (84 males, 25.5 ± 1.5 weeks gestation) within the whole cohort, 63 developed adverse outcomes, which were fewer for periods 2008–2011 (p = 0.013) and 2012–2015 (p = 0.035) compared with period 2006–2007 (adjusted for gestational age, Apgar scores and sex).ConclusionCareful bottom‐up revisions of protocols using iteration cycles, accounting for local settings, successfully improved the outcomes of preterm infants.