Infection Probability Score shows 59.4% specificity and 74.3% sensitivity for predicting infection in haematology–oncology patients
- Virginia Commonwealth University Medical Center
- Correspondence to
Gonzalo Bearman
PO Box 980019, Richmond, VA 23298-0019, USA; gbearman{at}mcvh-vcu.edu
Apostolpolou and colleagues assessed the predictive power of three systems: Infection Probability Score (IPS), Apache II and Karnovsky score for the onset of healthcare-associated infection in haematology–oncology patients. The sample consisted of 102 hospitalised patients, 32 of whom developed a healthcare-associated infection, the most frequent of which was a bloodstream infection (n=17). At a cut-off value of 10, the IPS demonstrated a healthcare-associated infection prediction sensitivity of 59.4% and a specificity of 74.3%, surpassing that of the APACHE II and Karnofsky scoring system.
Peres Bota1 initially developed the IPS to assess the probability of infection in critically ill patients. The IPS is simple, as the variables such as temperature, heart rate, respiratory rate, white blood cell count, C-reactive protein …
