Paediatric intensive care in Sweden : an epidemiological survey focusing on diagnostic panorama, outcome and factors influencing long-term mortality

Abstract: Background: In the 1990´s, studies of paediatric intensive care around the world had shown better outcome for children treated in PICUs compared to adult ICUs (AICU). In Sweden no nationwide data on children needing ICU care was present. Aim/Methods: To quantify the Swedish need for and outcome of intensive care for children, a retrospective multicenter cohort study was set up to include all children admitted to intensive care in from March 1998 to March 2001. The cohort was monitored for five years and survival data analyzed. (Study I). When analyzing the data, it was obvious that PIM2 score and a more extended data set to study factors involved in long-term mortality post PICU care was needed. A new three-year cohort was formed, this time with only PICU admissions from January 1, 2008 to December 31, 2010 (Study II). During the time of the study PIM2 score was not reported from AICUs. Arterial blood gas (ABG) is one among several variables included into the PIM2 score. We felt that to minimize unnecessary trauma, the routine use of ABG was unwarranted unless clinically indicated or for certain groups of children. We therefore studied a subset of the cohort to determine how PIM2 score predictability was influenced with or without ABG (Study III). It was noted that some of the children who died in the years after discharge, did so outside the PICU (Study II). We therefore performed an additional study of the 268 children whom died in the latter cohort to determine if limitation of medical treatment (LOMT) was the factor opposing PICU readmission for these children when turning fatally ill (Study IV). Results/Conclusions: We found that the outcome of intensive care for Swedish children was on par with international published data and 56% of the paediatric intensive care admissions were to AICUs. A 20-fold increased risk of death five years post PICU discharge was also found for the cohort (Study I). In the following cohort we found that having multiple admission (MADM) compared to single admissions (SADM) and/or a complex chronic conditions (CCC) significantly impaired the long-term outcome for five out of the seven different admission diagnosis groups (Study II). We also found that since Study I there was an increase in transfer from AICU to PICU from 65 to at least 278 children. In Study III we could show that the PIM2 score only becomes more accurate (although not significantly) if ABG is taken for the admission diagnostic group respiratory. In Study IV we in detail studied the 268 children that died in the cohort and could show that 123 (46%) died outside PICU. At the time of death 75% of them had a LOMT in place limiting readmission to PICU. Of the children not readmitted to PICU, 75% also had a CCC and 60% were males.