Beyond casemix: identifying an unpriced ICU cost driver in Victorian public hospitals

OBJECTIVE: Australia's National Pricing Model (NPM) adjusts public hospital funding for type, size, location, and patient complexity, but whether these adjustments adequately address structural cost variation at the jurisdictional level remains unknown. This study aimed to identify exogenous structural attributes associated with case-mix adjusted cost variation in Victorian public hospitals, and assess the adequacy of their compensation under activity-based funding. METHODS: A cross-sectional analysis of 1,934,946 inpatient episodes across 55 Victorian public hospital campuses (2024-25) was undertaken using linked administrative and cost data. Associations between 20 structural variables and cost per National Weighted Activity Unit (NWAU) were examined using Pearson/Spearman correlation, principal component analysis, and multivariable regression with HC3 heteroskedasticity-consistent standard errors. RESULTS: After adjustment for multiple comparisons, four variables were significantly associated with cost variation: weekend admissions (r = -0.524), after-hours admissions (r = -0.422), hospital remoteness (ρ = 0.328), and medical officer ratio (r = -0.305). In multivariable modelling, ICU non-Level 3 hours was the only consistent independent predictor of higher costs (P = 0.020). ICU non-Level 3 episodes cost $13,187 per NWAU - a $4063 gap above the National Efficient Price (62.8%) and a $73.0 million annual shortfall - with no current NPM adjustment. This contrasts with the more complex Level 3 ICU episodes, which have a NPM adjustment and no shortfall. CONCLUSIONS: Structural cost variation in Victorian hospitals follows a two-tiered pattern. First, after-hours, weekend, and teaching-intensity variables appear consistent with current NWAU pricing, despite being individually associated with cost variation. Second, ICU non-Level 3 hours represent a potentially unpriced structural cost driver warranting further investigation.