BackgroundCare bundles may reduce neonatal CLABSIs by 60% though it is often unclear if it is the intervention, the implementation, or both, that results in CLABSI reductions. This study aimed to investigate the implementation of a CLABSI care bundle in a UK neonatal intensive care unit (NICU). MethodsA focused ethnographic design was used underpinned by Normalisation Process Theory. Data collection included Normalisation MeAsure Development (NoMAD) surveys, observations of practice with dyadic think aloud interviews, and semi-structured interviews. Outcome and processes measures were collected. Data analysis used descriptive statistics and thematic analysis. ResultsThere was only partial bundle implementation, with minimal changes in survey scores and influences across individual, team and organisational levels moderating bundle adoption. Organisational culture sometimes undermined implementation and there were challenges relating to reinforcement and endorsement. ConclusionsThe introduction of a care bundle into a tertiary UK NICU did not result in sustained reductions in CLABSIs, which is one of few negative studies. Understanding the mechanisms by which an intervention works (or not) in specific contexts is important to optimise the delivery of evidence-based care.