Abstract
BACKGROUND: Specialized neurocritical care (NCC) improves outcomes in acute brain injury (ABI), but significant variability exists in practices across and hospitals within South Korea's developing national NCC system. This study aims to assess clinical variability among patients with ABI across six tertiary NCC units (NCCUs) in South Korea and evaluate center-specific effects on clinical outcomes. METHODS: A multicenter registry of patients with ABI admitted to NCCUs between April 2023 and April 2024 was analyzed. A descriptive analysis was conducted to evaluate demographic, clinical, and treatment characteristics across centers. Variability across centers was quantified using the average standardized mean difference (SMD) for key variables. Mixed-effects and fixed-effects models compared center-specific effects on 6- and 12-month functional outcomes (utility-weighted modified Rankin scale [mRS] score), in-hospital mortality, length of NCCU stay, and tracheostomy rates. RESULTS: Among 1,125 patients, 202 (18.2%) had aneurysmal subarachnoid hemorrhage, 478 (42.5%) had intracerebral hemorrhage, and 442 (39.3%) had traumatic brain injury. The median Glasgow Coma Scale (GCS) score was 13 (interquartile range 7-15). Notable differences (SMD > 0.2) were observed in premorbid mRS scores, initial clinical severity (e.g., GCS, pupillary response), treatment practices (e.g. intracranial pressure monitoring, vasospasm prophylaxis), and outcomes (e.g., 6-month mRS score). Hospital-specific effects did not significantly influence most outcomes; mixed-effects models showed no significant improvement in model fit for 6-month mRS scores (P = 0.78), in-hospital mortality (P = 0.99), length of NCCU stay (P = 0.12), and tracheostomy rates (P = 0.11), except for the 12-month mRS score (P = 0.01). CONCLUSIONS: Significant variability exists among patients with ABI across NCCUs in South Korea. Despite these differences, center-specific effects did not significantly influence key clinical outcomes closely related to NCCU care, suggesting that variability in outcomes may be more attributable to patient-level factors.