CLINICAL PREDICTIVE MODELING IN CHILDREN

1 Benjamin D, DeLong E 2 Steinbach W, Cotten C 3 Walsh T, 4 Clark R

Author address: 

1 Duke University Clinical Research, Durham, United States 2 Duke University Pediatrics, Durham, United States 3 National Cancer Institute, Bethesda, MD, United StatesDurham, United States 4 Pediatrix Medical Group, Inc

Abstract: 

Background: Neonatal candidemia is difficult to diagnose and often fatal. Empirical antifungal therapy is associated with improved survival in neonates and patients with fever and neutropenia. Although guidelines for empirical therapy exist for patients with fever and neutropenia, these do not exist for neonates. Design: Multi-center retrospective cohort study of neonatal intensive care unit patients (N=6,172). We evaluated all blood cultures (N=21,233) from neonates older than day of life 3 with birthweight 28 weeks. Children with a candidemia score >2 points were classified as having a "œpositive" candidemia score. The candidemia score had a sensitivity of 85% and specificity of 47%. Conclusions: We developed a clinical predictive model for neonatal candidemia with high sensitivity and reasonable specificity for candidemia. Based on our model, when a physician obtains a blood culture, all neonates
2003

abstract No: 

none

Full conference title: 

13th Annual Focus on Fungal Infections
    • FFI 13th (2003)