We read with interest the paper from McCredie et al in Critical Care Medicine. The authors retrospectively analysed data from 9773 patients from 134 intensive care units (ICUs), obtained between 2011 and 2013, extracted from a quality improvement dataset (the American College of Surgeons Trauma Quality Improvement Program; ACS TQIP) They used these data to seek associations between structures and processes of care and in-hospital mortality in traumatic brain injury (TBI). Their analysis showed that clinical outcome was not related to ICU structure (dedicated neurocritical care units vs. general ICUs), but was improved by the presence of standardised management protocols. We would make the following comments.