592

Abstract

Introduction: Others have demonstrated modest inverse relationships between PICU patient volume and severity-adjusted mortality rates (1,2). We previously were unable to demonstrate a clear relationship between overall PICU volume and severity-adjusted mortality (3). Methods: The VPS database (VPS, LLC) was queried for all patients discharged from September 2009 through March 2012 with valid pediatric index of mortality 2 (PIM2) and PRISM3 scores. Anonymized data received included ICU mortality, PIM2 and PRISM3 score. Average PICU volume/quarter (VOL) was calculated as total discharges divided by the number of valid quarters (VPS sites submit data quarterly) per PICU. VOL was then divided by 100 (VOL-100) to assess the impact per 100 discharges per quarter. The population was divided into quintiles based on VOL-100. Several multiple logistic regression models were performed to assess the impact of VOL-100 on severity-adjusted mortality for the entire sample and within each quintile (reported as odds ratios (OR with 95% CI's). This study received a waiver as non-human research from the IRB at CHLA. Results: From 113 PICU's, 194,400 patients were studied, with an overall ICU mortality rate of 2.6%. VOL ranged from 32 to 881; the mean VOL was 269. There was a trend towards lower mortality rates as VOL-100 increased; OR.0.98 (0.97, 1.00; p = 0.08). However, the OR for VOL-100 was 1.06 (1.04, 1.08) when adjusted with PIM2 and 1.03 (1.00, 1.06) with PRISM3. In the quintile analysis, there was an increased risk of mortality per unit of VOL-100 within the lowest volume quintile; unadjusted OR 1.76 (1.48, 2.08), that remained significant after adjustment with either PIM2 (OR 1.62; 1.34, 1.95) or PRISM3 (OR 1.29; 1.04, 1.61). In the highest volume quintile, VOL-100 was associated with reduced mortality; unadjusted OR 0.88 (0.81, 0.95) that was not significant once adjusted with either PIM2 (OR 0.98; 0.89, 1.08) or PRISM3 (OR 0.92; 0.82, 1.02). Conclusions: In this preliminary analysis, the relationship between PICU volume and mortality appears to vary with unit size. In the smallest volume units, mortality increases with volume, even after severity-of-illness adjustment. There does not appear to be a relationship between volume and outcome in the highest volume units after adjusting for severity of illness. 1. Tilford et al. Pediatrics 2000; 106:289-294. 2. Marcin et al. Pediatr Crit Care Med 2005; 6:136-141. 3. Markovitz et al. Crit Care Med 2009; 37(12):A356