Using Time Series Analysis to Predict Cardiac Arrest in a PICU.

OBJECTIVES: To build and test cardiac arrest prediction models in a PICU, using time series analysis as input, and to measure changes in prediction accuracy attributable to different classes of time series data. DESIGN: Retrospective cohort study. SETTING: Thirty-one bed academic PICU that provides care for medical and general surgical (not congenital heart surgery) patients. SUBJECTS: Patients experiencing a cardiac arrest in the PICU and requiring external cardiac massage for at least 2 minutes. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: One hundred three cases of cardiac arrest and 109 control cases were used to prepare a baseline dataset that consisted of 1,025 variables in four data classes: multivariate, raw time series, clinical calculations, and time series trend analysis. We trained 20 arrest prediction models using a matrix of five feature sets (combinations of data classes) with four modeling algorithms: linear regression, decision tree, neural network, and support vector machine. The reference model (multivariate data with regression algorithm) had an accuracy of 78% and 87% area under the receiver operating characteristic curve. The best model (multivariate + trend analysis data with support vector machine algorithm) had an accuracy of 94% and 98% area under the receiver operating characteristic curve. CONCLUSIONS: Cardiac arrest predictions based on a traditional model built with multivariate data and a regression algorithm misclassified cases 3.7 times more frequently than predictions that included time series trend analysis and built with a support vector machine algorithm. Although the final model lacks the specificity necessary for clinical application, we have demonstrated how information from time series data can be used to increase the accuracy of clinical prediction models.

Procalcitonin in children with Escherichia coli O157:H7 associated hemolytic uremic syndrome.

Shiga toxin producing Escherichia coli (STEC) are noninvasive enteric pathogens that may cause hemorrhagic colitis (HC) and diarrhea-associated hemolytic uremic syndrome (D+ HUS). We hypothesized that development of D+ HUS is associated with increased serum procalcitonin (PCT) levels. PCT was measured by an immunoluminometric assay in 113 patients. Concentrations of PCT were different in normal controls, disease control groups (rotavirus enteritis, HC due to non-STEC pathogens, chronic renal failure), and children with uncomplicated O157:H7 HC or D+ HUS. Children with D+ HUS showed higher PCT levels than those with uncomplicated O157:H7 HC, and increased concentrations were noted in cases requiring peritoneal dialysis. Severely increased concentrations were observed in children with D+ HUS on d 5 or 6 after the onset of enteritis, whereas serial measurements in those with uncomplicated O157:H7 HC remained within the normal range throughout the first week of illness. PCT was correlated with serum concentrations of lipopolysaccharide (LPS)-binding protein and serum levels of alanine aminotransferase. Using two separate sets of real-time PCR primers, we were unable to detect elevated PCT mRNA transcripts in nonadherent undifferentiated (monocytic) or differentiated (macrophage-like) THP-1 cells stimulated with purified Shiga toxin-1 and/or LPS. Our data show that serum levels of PCT are associated with the severity of illness in children with D+ HUS. Further studies are needed to determine the role of PCT in the pathogenesis of thrombotic microangiopathy associated to childhood D+ HUS.