Volume 13
Number 0
Feb 2009
Back to issue
Other
PHJ van der Voort, F Bakhshi-Raiez, DW de lange, RJ Bosman, E de Jonge, H Joore, R de Waal, RMJ Wesselink, G van Berkel, R van Maanen, NF de Keizer

Introduction The NICE registry contains data from individual intensive care admissions concerning severity of illness and outcome. This study was undertaken to analyze the trends in time related to crude mortality and severity of illness corrected mortality. Methods Data on all patients admitted from January 1 1999 to January 1 2008 were used. Patients who did not fulfil the inclusion criteria for the SAPS II prediction model were excluded. To analyze trends in time for individual ICUs, this study uses data from ICUs that participated in the NICE registry for a minimum of 5 years between January 1, 2001 and January 1, 2008. Repeated cross-sectional data analysis was performed for Standardized Mortality Ratio (SMR), and longitudinal data analysis using Variable Live Adjusted Displays (VLAD both on individual and national level were used. Results Overall, 243 182 patients were included in the NICE registry. Of these, 144 696 met the SAPS II inclusion criteria and were included in this study. In the non-cardiac surgery population, the SAPS II model was customized on 9227 admissions from 13 ICUs between 1999 and 2000 to obtain a baseline Dutch SMR of 1.0 for comparison of subsequent years. In individual ICUs, the SAPS II model was customized on 15 863 admissions from 20 ICUs that participated in NICE for a minimum of five years between 2001 and 2002 to obtain a baseline SMR of 1.0 for comparison of subsequent years. From 2001 to 2007 in the non-cardiac surgery population, a significant decrease in SMR was shown from 1.0 tot 0.83 (p=0.023). This decrease was most significant in patients after elective surgery (p=0.01). Seven (25%) individual ICUs showed a significant decrease in their SMR. The VLAD curve for individual ICUs showed a highly variable course. Conclusion Over time, the NICE registry showed a significant improvement in SMR for all patient groups nationwide, and for 7 individual ICUs. Longitudinal analysis for individual ICUs by VLAD curve shows a variable course between ICUs and within an individual ICU over time. This information can be used for quality improvement initiatives both locally and nationwide


Back to issue - Download PDF