Objectives: To assess the association between ED occupancy and relevant outcomes including ED waiting times, rates of admission and representation and length of stay when hospitalised. Methods: Retrospective study of all ED presentations by New South Wales (NSW), Australia, residents to 15 NSW public, principal referral or paediatric specialist hospitals between 1 January to 31 December 2015 (N = 935 282). ED data were linked longitudinally (to ED data) and cross-sectionally to hospital admissions data. An ED-system measure of occupancy was assigned to each ED record. The study outcomes were ED waiting time, admission to hospital, 28 day representation, and length of stay (LOS) when admitted. Outcomes were analysed using univariate analyses and multivariable general linear and binary logistic regression models. Results: Increased ED occupancy was associated with increased ED waiting times, particularly at lowbaseline occupancy (e.g. rate ratio = 2.22, 95% confidence interval [CI] [2.10-2.35], for non-urgent triaged patients). However, results were conditional on triage category, such that estimated effects were smaller or not significant in emergency and resuscitation triaged patients (e.g. rate ratio = 1.59, 95% CI [1.52-1.65], for emergency patients). ED occupancy only showed small or no associations with admission to hospital, 28 day representation and LOS when admitted. Conclusions: Higher ED occupancy was associated with increased waiting times conditional on triage category and baseline occupancy. Collectively, the results show that NSW principal referral EDs are robust, and are currently capable of handling variation in occupancy by prioritising treatment for the most urgent patients.